Java on Chase the Devil

A Bachelier Normal Implied Volatility Solver

Sun, 07 Jun 2026 03:43:00 +0000

Around 2014, I proposed a few Bachelier implied volatility “solvers”. The first one was inspired by Steven G. Johnson’s Faddeeva package for the complex error function: it used a piecewise Chebyshev polynomial representation to have a near machine accurate representation of the Bachelier implied volatility. Why did I put “solver” in quotes? Because the problem can be reduced to a 1D function representation (as originally shown by Choi, Kim and Kwak in 2009). Then, I was encouraged by Gary Kennedy to find a simpler rational function approximation, I used 3 or 4 rational functions to cover the full range of the Bachelier implied volatility function.

The Role of Erfcx and Erfc in the Black-Scholes Formula

Sat, 30 May 2026 03:43:00 +0000

This note compares special-function choices at the level that matters to the implied volatility solvers: the normalized Black beta price, not standalone erfcx(x).

For x = log(F/K) <= 0 and total volatility s = sigma sqrt(T), the beta-space OTM call price is

B(x, s) = 0.5 * (exp(x/2) * erfc(q1) - exp(-x/2) * erfc(q2))
q1 = -(x/s + s/2) / sqrt(2)
q2 = -(x/s - s/2) / sqrt(2)

The corresponding forward-normalized OTM-call price is c = B(x, s) / sqrt(ex), with ex = exp(x).

Which erfcx?

Sun, 24 May 2026 03:43:00 +0000

The use of erfcx instead of direct erfc or CDF in a Black-Scholes implied volatility solver leads to gain in accuracy and performance in general. But which erfcx should we use?

This note compares practical erfcx implementations for Rust implied volatility solvers:

Commons: the local Rust port of Apache Commons Numbers BoostErf.erfcx
Cody: the Cody rational approximation exposed by the jaeckel = 0.2.0 crate during this comparison, originally coming from Netlib SLATEC library.
Johnson: Steven G. Johnson’s Faddeeva implementation, as exposed by errorfunctions = 0.2.0, faddeeva-sys = 0.1.0, and Julia SpecialFunctions.erfcx(::Float64)

The Commons-style implementation is a local Rust port of the Apache Commons Numbers/BoostErf approach. It computes the small central region through the local erf rational approximation and uses exp_m1(x*x) to avoid cancellation in exp(x^2) * erfc(x) near zero. For positive inputs it uses several rational subintervals: a central polynomial/rational erf branch, erfc-style rational branches around 0.5..1.5, 1.5..2.5, and 2.5..4, and a rational function taken from Rational Chebyshev approximations for the error function by W. J. Cody, Math. Comp., 1969, PP. 631-638 beyond that. For very large positive x, it returns 1 / (sqrt(pi) * x). For negative x, it uses the reflected form with a split-square exponential, returns Inf below the finite overflow cutoff, and otherwise computes 2 exp(x^2) - erfcx(|x|) in a way that preserves more bits near the overflow boundary.

A Faster Monotone Implied Volatiltty Solver

Sat, 23 May 2026 03:43:00 +0000

Choi, Huh and Su have a very good paper entitled Tighter uniform bounds for Black–Scholes implied volatility and the applications to root-finding. What’s particularly great is that it gives both a decent lower bound and a proof a monotone convergence using Newton’s method starting from this lower bound.

The industry standard for solving the Black-Scholes implied volatility is Peter Jäckel Let’s be rational. I was wondering if one could not create a robust mathematically backed solver at least as fast and accurate as Jäckel’s algorithm.

Senior Developers Don't Know OO Anymore

Thu, 08 Mar 2018 20:56:42 +0100

It has been a while since the good old object-oriented (OO) programming is not trendy anymore. Functional programming or more dynamic programming (Python-based) have been the trend, with an excursion in template based programming for C++ guys. Those are not strict categories: Python can be used in a very OO way, but it’s not how it is marketed or considered by the community.

Recently, I have seen some of the ugliest refactoring in my life as a programmer, done by someone with at least 10 years of experience programming in Java. It is a good illustration because the piece of code is particularly simple (although I won’t bother with implementation details). The original code was a simple boolean method on an object such as

Modern Programming Language for Monte-Carlo

Sat, 18 Apr 2015 22:58:00 +0000

A few recent programming languages sparked my interest:

Julia because of the wide coverage of mathematical functions, and great attention to quality of the implementations. It has also some interesting web interface.
Dart: because it’s a language focused purely on building apps for the web, and has a supposedly good VM.
Rust: it’s the latest fad. It has interesting concepts around concurrency and a focus on being low level all the while being simpler than C.

I decided to see how well suited they would be on a simple Monte-Carlo simulation of a forward start option under the Black model. I am no expert at all in any of the languages, so this is a beginner’s test. I compared the runtime for executing 16K simulations times a multiplier.

On the Role of Static Types and Generic Types on Productivity

Sun, 29 Jun 2014 10:40:00 +0000

Most developers have strong opinions on dynamic types programming languages vs static types programming languages. The former is often assumed to be good for small projects/prototyping while the later better for bigger projects. But there is a surprisingly small number of studies to back those claims.

One such study is "An experiment about static and dynamic type systems: doubts about the positive impact of static type systems on development time" and came to the conclusion that on a small project, static typing did not decrease programming time, and actually increased debugging time. However 4 years later, "An empirical comparison of static and dynamic type systems on API usage in the presence of an IDE: Java vs. groovy with eclipse" shows that a developer is 2x more productive with Java than with Groovy using an unknown API. This contrasts a bit (but does not contradict) with their previous study "Static Type Systems (Sometimes) have a Positive Impact on the Usability of Undocumented Software: An Empirical Evaluation" that showed Groovy to be more productive on small projects. One problem is that all these studies stem from the same person.

It's more interesting to look at generic types vs raw types use, where even less studies have been done. "Do developers benefit from generic types?: an empirical comparison of generic and raw types in java" concludes that generic types do not provide any advantages to fix typing errors, hardly surprising in my opinion. Generic types (especially with type erasure as in Java) is the typical idea that sounds good but that in practice does not really help: it makes the code actually more awkward to read and tend to make developers too lazy to create new classes that would often be more appropriate than a generic type (think Map<String,List<Map<String, Date>>>).

throws Exception

Tue, 27 May 2014 10:49:00 +0000

There was a big debate at work around Exception declaration in a Java API. I was quite surprised that such an apparently simple subject could end up being so controversial. The controversy was around the choice of declaring in the interfaces:

void myMethod() throws Exception

instead of

void myMethod() throws MyAPIException
void myMethod() throws MyAPIRuntimeException
void myMethod()

where MyAPI represents either a generic API related exception or a specific exception related to the method in question.

The choice of "throws Exception" did not even occur to me as a possibility, but after some digging, I found that some relatively famous libraries actually followed that principle at one point, for example Apache Struts 1.x or Spring MVC.

More modern libraries, like Google Guava, commons-math 3.x, Struts 2.x generally favor MyAPIRuntimeException where MyAPI is actually context-specific. Some old popular libraries declare a checked Exception, for example the HibernateException in Hibernate.

This seems to be a recurring subject on Stackoverflow:
Stackoverflow - Java interface throws Exception best practice
Stackoverflow - What to put in the throws clause of an interface method

But those are quite poor in terms of explanations. The best comments on this subjects are from:
Anders Hejlsberg (C#, Delphi, Turbo Pascal creator) - The Trouble with Checked Exceptions
James Gosling (Java creator) - Failure and Exceptions

This comment from Anders is particularly acute:
"To work around this requirement, people do ridiculous things. For example, they decorate every method with, "throws Exception." That just completely defeats the feature, and you just made the programmer write more gobbledy gunk. That doesn't help anybody. "

Today I believe the API in question declares "throws Exception"...

5 Minutes of Xtend

Tue, 08 Apr 2014 17:37:00 +0000

There is a relatively new JVM based language, Xtend. Their homepage says “JAVA 10, TODAY!”, so I thought I would give it a try, I was especially interested in operator overloading support, and the fact that it compiles to Java code, not Java byte code.

Unfortunately, after 5 minutes with it, and pasting some non Java code in an xtend file, Eclipse hangs forever, even on restart. After creating another workspace, just to trash the new workspace a similar way. This is quite incredible for a nearly 2 years old project, on eclipse.org.

Julia and the Cumulative Normal Distribution

Tue, 13 Aug 2013 15:52:00 +0000

I just stumbled upon Julia, a new programming language aimed at numerical computation. It’s quite new but it looks very interesting, with the promise of C like performance (thanks to LLVM compilation) with a much nicer syntax and parallelization features.Out of curiosity, I looked at their cumulative normal distribution implementation. I found that the (complimentary) error function (directly related to the cumulative normal distribution) algorithm relies on an algorithm that can be found in the Faddeeva library. I had not heard of this algorithm or this library before, but the author, Steven G. Johnson, claims it is faster and as precise as Cody & SLATEC implementations. As I previously had a look at those algorithms and was quite impressed by Cody’s implementation.The source of Faddeeva shows a big list (100) of Chebychev expansions for various ranges of a normalized error function. I slightly modified the Faddeva code to compute directly the cumulative normal distribution, avoiding some exp(-xx)exp(xx) calls on the way.Is it as accurate? I compared against a high precision implementation as in my previous test of cumulative normal distribution algorithms. And after replacing the exp(-xx) with Cody’s trick to compute it with higher accuracy, here is how it looks (referenced as “Johnson”).I also measured performance on various ranges, and found out that this Johnson algorithm is around 2x faster than Cody (in Scala) and 30% faster than my optimization of Cody (using a table of exponentials for Cody’s trick).

Octave vs Scilab for PDEs in Finance

Tue, 30 Jul 2013 12:10:00 +0000

I was used to Scilab for small experiments involving linear algebra. I also like some of Scilab choices in algorithms: for example it provides PCHIM monotonic spline algorithm, and uses Cody for the cumulative normal distribution.

Matlab like software is particularly well suited to express PDE solvers in a relatively concise manner. To illustrate some of my experiments, I started to write a Scilab script for the Arbitrage Free SABR problem. It worked nicely and is a bit nicer to read than my equivalent Scala program. But I was a bit surprised by the low performance.

The CUDA Performance Myth II

Fri, 12 Jul 2013 15:23:00 +0000

This is a kind of following to the CUDA performance myth. There is a recent news on the java concurrent mailing list about SplittableRandom class proposed for JDK8. It is a new parallel random number generator a priori usable for Monte-Carlo simulations.

It seems to rely on some very recent algorithm. There are some a bit older ones: the ancestor, L’Ecuyer MRG32k3a that can be parallelized through relatively costless skipTo methods, a Mersenne Twister variant MTGP, and even the less rigourous XorWow popularized by NVidia CUDA.

Simple 'Can Scala Do This?' Questions

Tue, 11 Jun 2013 00:28:00 +0000

Today, a friend asked me if Scala could pass primitives (such as Double) by reference. It can be useful sometimes instead of creating a full blown object. In Java there is commons lang MutableDouble. It could be interesting if there was some optimized way to do that.

One answer could be: it’s not functional programming oriented and therefore not too surprising this is not encouraged in Scala.

Then he wondered if we could use it for C#.

SABR with the new Hagan PDE Approach

Tue, 28 May 2013 15:56:00 +0000

At a presentation of the Thalesians, Hagan has presented a new PDE based approach to compute arbitrage free prices under SABR. This is similar in spirit as Andreasen-Huge, but the PDE is directly on the density, not on the prices, and there is no one-step procedure: it's just like a regular PDE with proper boundary conditions.

I was wondering how it compared to Andreasen Huge results.

My first implementation was quite slow. I postulated it was likely the Math.pow function calls. It turns out they could be reduced a great deal. As a result, it's now quite fast. But it would still be much slower than Andreasen Huge. Typically, one might use 40 time steps, while Andreasen Huge is 1, so it could be around a 40 to 1 ratio. In practice it's likely to be less than 10x slower, but still.

While looking at the implied volatilities I found something intriguing with Andreasen Huge: the implied volatilities from the refined solution using the corrected forward volatility look further away from the Hagan implied volatilitilies than without adjustment, and it's quite pronounced at the money.

Interestingly, the authors don't plot that graph in their paper. They plot a similar graph of their own closed form analytic formula, that is in reality used to compute the forward volatility. I suppose that because they calibrate and price through their method, they don't really care so much that the ATM prices don't match Hagan original formula.

We can see something else on that graph: Hagan PDE boundary is not as nice as Andreasen Huge boundary for high strikes (they use a Hagan like approx at the boundaries, this is why it crosses the Hagan implied volatilities there).

If we use a simple option gamma = 0 boundary in Andreasen Huge, this results in a very similar shape as the Hagan PDE. This is because the option price is effectively 0 at the boundary.
Hagan chose a specifically taylored Crank-Nicolson scheme. I was wondering how it fared when I reduced the number of time-steps.

The answer is: not good. This is the typical Crank-Nicolson issue. It could be interesting to adapt the method to use Lawson-Morris-Goubet or TR-BDF2, or a simple Euler Richardson extrapolation. This would allow to use less time steps, as in practice, the accuracy is not so bad with 10 time steps only.

What I like about the Hagan PDE approach is that the implied vols and the probability density converge well to the standard Hagan formula, when there is no negative density problem, for example for shorter maturities. This is better than Andreasen Huge, where there seems to be always 1 vol point difference. However their method is quite slow compared to the original simple analytic formula.

Update March 2014 - I have now a paper around this "Finite Difference Techniques for Arbitrage Free SABR"

A Fast Exponential Function in Java

Fri, 19 Apr 2013 16:48:00 +0000

In finance, because one often dicretize the log process instead of the direct process for Monte-Carlo simulation, the Math.exp function can be called a lot (millions of times for a simulation) and can be a bottleneck. I have noticed that the simpler Euler discretization was for local volatility Monte-Carlo around 30% faster, because it avoids the use of Math.exp.

Can we improve the speed of exp over the JDK one? At first it would seem that the JDK would just call either the processor exp using an intrinsic function call and that should be difficult to beat. However what if one is ok for a bit lower accuracy? Could a simple Chebyshev polynomial expansion be faster?

Out of curiosity, I tried a Chebyshev polynomial expansion with 10 coefficients stored in a final double array. I computed the coefficient using a precise quadrature (Newton-Cotes) and end up with 1E-9, 1E-10 absolute and relative accuracy on [-1,1].

Here are the results of a simple sum of 10M random numbers:

0.75s for Math.exp sum=1.7182816693332244E7
0.48s for ChebyshevExp sum=1.718281669341388E7
0.40s for FastMath.exp sum=1.7182816693332244E7

So while this simple implementation is actually faster than Math.exp (but only works within [-1,1]), FastMath from Apache commons maths, that relies on a table lookup algorithm is just faster (in addition to being more precise and not limited to [-1,1]).

Of course if I use only 5 coefficients, the speed is better, but the relative error becomes around 1e-4 which is unlikely to be satisfying for a finance application.

0.78s for Math.exp sum=1.7182816693332244E7
0.27s for ChebyshevExp sum=1.718193001875838E7
0.40s for FastMath.exp sum=1.7182816693332244E7

Root finding in Lord Kahl Method to Compute Heston Call Price (Part II)

Thu, 11 Apr 2013 16:29:00 +0000

In my previous post, I explored the Lord-Kahl method to compute the call option prices under the Heston model. One of the advantages of this method is to go beyond machine epsilon accuracy and be able to compute very far out of the money prices or very short maturities. The standard methods to compute the Heston price are based on a sum/difference where both sides are far from 0 and will therefore be limited to less than machine epsilon accuracy even if the integration is very precise.

Root finding in Lord Kahl Method to Compute Heston Call Price

Tue, 09 Apr 2013 19:49:00 +0000

I just tried to implement Lord Kahl algorithm to compute the Heston call price. The big difficulty of their method is to find the optimal alpha. That’s what make it work or break. The tricky part is that the function of alpha we want to minimize has multiple discontinuities (it’s periodic in some ways). This is why the authors rely on the computation of an alpha_max: bracketing is very important, otherwise your optimizer will jump the discontinuity without even noticing it, while you really want to stay in the region before the first discontinuity.

From Double Precision Normal Density to Double Precision Cumulative Normal Distribution

Tue, 02 Apr 2013 14:24:00 +0000

Marsaglia in his paper on Normal Distribution made the same mistake I initially did while trying to verify the accuracy of the normal density.In his table of values comparing the true value computed by Maple for some values of x to the values computed by Sun or Ooura erfc, he actually does not really use the same input for the comparison. One example is the last number: 16.6. 16.6 does not have an exact representation in double precision, even though it is displayed as 16.6 because of the truncation at machine epsilon precision. Using Python mpmath, one can see that:>>> mpf(-16.6)mpf(’-16.6000000000000014210854715202004’)This is the more accurate representation if one goes beyond double precision (here 30 digits). And the value of the cumulative normal distribution is:>>> ncdf(-16.6)mpf(‘3.4845465199503256054808152068743e-62’)It is different from:>>> ncdf(mpf("-16.6"))mpf(‘3.48454651995040810217553910503186e-62’)where in this case it is really evaluated around -16.6 (up to 30 digits precision). Marsaglia gives this second number as reference. But all the other algorithms will actually take as input the first input. It is more meaningful to compare results using the exact same input. Using human readable but computer truncated numbers is not the best. The cumulative normal distribution will often be computed using some output of some calculation where one does not have an exact human readable input.The standard code for Ooura and Schonfelder (as well as Marsaglia) algorithms for the cumulative normal distribution don’t use Cody’s trick to evaluate the exp(-xx). This function appears in all those implementations because it is part of the dominant term in the usual expansions. Out of curiosity, I replaced this part with Cody trick. For Ooura I also made minor changes to make it work directly on the CND instead of going through the error function erfc indirection. Here are the results without the Cody trick (except for Cody):and with it:All 3 algorithms are now of similiar accuracy (note the difference of scale compared to the previous graph), with Schonfelder being a bit worse, especially for x >= -20. If one uses only easily representable numbers (for example -37, -36,75, -36,5, …) in double precision then, of course, Cody trick importance won’t be visible and here is how the 3 algorithms would fare with or without Cody trick:Schonfelder looks now worse than it actually is compared to Cody and Ooura.To conclude, if someone claims that a cumulative normal distribution is up to double precision accuracy and it does not use any tricks to compute exp(-xx), then beware, it probably is quite a bit less than double precision.

Scala is Mad (part 2)

Wed, 13 Feb 2013 16:20:00 +0000

I still did not abandon Scala despite my previous post, mainly because I have already quite a bit of code, and am too lazy to port it. Furthermore the issues I detailed were not serious enough to motivate a switch. But these days I am more and more fed up with Scala, especially because of the Eclipse plugin. I tried the newer, the beta, and the older, the stable, the conclusion is the same. It’s welcome but:code completion is not great compared to Java. For example one does not seem to be able to see the constructor parameters, or the method parameters can not be automatically populated.the plugin makes Eclipse very slow. Everything seems at least 3-5x slower. On the fly compilation is also much much slower than Java’s.It’s nice to type less, but if overall writing is slower because of the above issues, it does not help. Beside curiosity of a new language features, I don’t see any point in Scala today, even if some of the ideas are interesting. I am sure it will be forgotten/abandoned in a couple of years. Today, if I would try a new language, I would give Google Go a try: I don’t think another big language can make it/be useful on the JVM (beside a scripting kind of language, like JavaScript or Jython).Google Go focuses on the right problem: concurrency. It also is not constrained to JVM limitation (on the other side one can not use a Java library - but open source stuff is usually not too difficult to port from one language to another). It has one of the fastest compilers. It makes interesting practical choices: no inheritance.

Scala is Mad

Wed, 12 Dec 2012 16:07:00 +0000

I spent quick a bit of time to figure out why something that is usually simple to do in Java did not work in Scala: Arrays and ArrayLists with generics.

For some technical reason (type erasure at the JVM level), Array sometimes need a parameter with a ClassManifest !?! a generic type like [T :< Point : ClassManifest] need to be declared instead of simply [T :< Point].

And then the quickSort method somehow does not work if invoked on a generic… like quickSort(points) where points: Array[T]. I could not figure out yet how to do this one, I just casted to points.asInstanceOf[Array[Point]], quite ugly.

Scala Again

Mon, 06 Feb 2012 17:52:00 +0000

I am trying Scala again. Last time, several years ago, I played around with it as a web tool, combining it with a Servlet Runner like Tomcat. This time, I play around with it for some quantitative finance experiments.

Why Scala? It still seem the most advanced alternative to Java on the JVM, and the mix of functional programming and OO programming is interesting. Furthermore it goes quite far as it ships with its own library. I was curious to see if I could express some things better with Scala.

Here are my first impressions after a week:

I like the object keyword. It avoids the messy singleton pattern, or the classes with many static methods. I think it makes things much cleaner to not use static at all but distinguish between object & class.
I like the Array[Double], and especially ArrayBuffer[Double]. Finally we don't have to worry between the Double and double performance issues.
I was a bit annoyed by a(i) instead of a[i] but it makes sense. I wonder if there is a performance implication for arrays, hopefully not.
I like the real properties, automatic getter/setter: less boilerplate code, less getThis(), setThat(toto).
Very natural interaction with Java libraries.
I found a good use of case classes (to my surprise): typically an enum that can have some well defined parameters, and that you don't want to make a class (because it's not). My use case was to define boundaries of a spline.
I love the formatter in the scala (eclipse) IDE. Finally a formatter in eclipse that does not produce crap.

Now things I still need time to get used to:

member variable declared implicitly in the constructor. I first made the mistake (still?) to declare some variables twice.
I got hit by starting a line with a + instead of ending with a +. It is dangerous, but it certainly makes the code more consistent.
Performance impacts: I will need to take a look at the bytecode for some scala constructs to really understand the performance impact of some uses. For example I tend to use while loops instead of for comprehension after some scary post of the Twitter guys about for comprehension. But at first, it looks as fast as Java.
I wrote my code a bit fast. I am sure I could make use of more Scala features.
The scala IDE in eclipse 3.7.1 has known issues. I wish it was a bit more functional, but it's quite ok (search for references works, renaming works to some extent).
Scala unit tests: I used scala tests, but it seems a bit funny at first. Also I am not convinced by the syntax that avoid method names and prefer test("test name"). It makes it more difficult to browse the source code.

Some things they should consider:

Integrate directly a Log API. I just use SLF4J without any scala wrapper, but it feels like it should be part of the standard API (even if that did not work out so well for Sun).
Double.Epsilon is not the machine epsilon: very strange. I found out somewhere else there was the machine epsilon, don't remember where because I ended up just making a small object.
Unit tests should be part of the standard API.

Overall I found it quite exciting as there are definitely new ways to solve problems. It was a while since I had been excited with actual coding.

exp(y*log(x)) Much Faster than Math.pow(x,y)

Fri, 08 Apr 2011 23:03:00 +0000

Today I found out that replacing Math.pow(x,y) by Math.exp(yMath.log(x))* made me gain 50% performance in my program. Of course, both x and y are double in my case. I find this quite surprising, I expected better from Math.pow.

SIMD and Mersenne-Twister

Sat, 05 Feb 2011 13:18:00 +0000

Since 2007, there is a new kind of Mersenne-Twister (MT) that exploits SIMD architecture, the SFMT. The Mersenne-Twister has set quite a standard in random number generation for Monte-Carlo simulations, even though it has flaws.

I was wondering if SFMT improved the performance over MT for a Java implementation. There is actually on the same page a decent Java port of the original algorithm. When I ran it, it ended up slower by more than 20% than the classical Mersenne-Twister (32-bit) on a 64-bit JDK 1.6.0.23 for Windows.

The CUDA Performance Myth

Mon, 03 Jan 2011 16:07:00 +0000

There is an interesting article on how to generate efficiently the inverse of the normal cumulative distribution on the GPU. This is useful for Monte-Carlo simulations based on normally distributed variables.

Another result of the paper is a method (breakless algorithm) to compute it apparently faster than the very good Wichura’s AS241 algorithm on the CPU as well keeping a similar precision. The key is to avoid branches (if-then) at the cost of not avoiding log() calls. As the algorithm is very simple, I decided to give it a try in Java.

Another Look at Java Matrix Libraries

Mon, 29 Nov 2010 12:45:00 +0000

A while ago, I was already looking for a good Java Matrix library, complaining that there does not seem any real good one where development is still active: the 2 best ones are in my opinion Jama and Colt.

Recently I tried to price options via RBF (radial basis functions) based on TR-BDF2 time stepping. This is a problem where one needs to do a few matrix multiplications and inverses (or better, LU solve) in a loop. The size of the matrix is typically 50x50 to 100x100, and one can loop between 10 and 1000 times.

Java enum Is Evil

Thu, 12 Aug 2010 17:32:00 +0000

Before Java 1.5, I never really complained about the lack of enum keyword. Sure the old enum via class pattern was a bit verbose at first (N.B.: Java 1.5 enums can also be verbose once you start adding methods to them). But more importantly, you would often use the table lookup pattern in combination.

The problem with Java 1.5 enum is that it is not Object-Oriented. You can't extend an enum, you can't add an element in an existing enum. Many will say "but that's what enum is for, a static list of things". In my experience, the list of things often changes with time, or needs to be extended at one point. Furthermore, most people (including me when I am very lazy) end up writing switch statements on enum values. Enum promotes bad programming practices.

Think twice about using enum, this is often not what you want.

A Very Interesting Feature of Scala

Sat, 07 Aug 2010 12:35:00 +0000

I tried Scala a few years ago. There are several good ideas in it, but I found the language to be a bit too complicated to master. But I recently stubbled upon a paper on Scala generics that might change my mind about using Scala.

Scala Generics used to work in a similar way as Java Generics: via type erasure. One main reason is compatibility with Java, another is that C++ like templates make the code base blow up. Scala Generics offered some additional behavior (the variance/covariance notion). C++ templates, however, have some very interesting aspects: one is that everything is done at compile time, the other is performance. If the generics are involved in any kind of computation intensive task, all the Java type conversion will create a significant overhead.

Now Scala has @specialized (since Scala 2.8). Annotating a generic type with @specialized will generate code. One has the choice to accept the performance penalty or to get all the performance but accept the code blow up. I think this is very useful.

If you read the paper you will see that the performance implications of this are not always small.

UPDATE: I thank the readers for pointing that this work only with primitive types to avoid autoboxing. It is still valuable but less than I first thought.

Diffusion Limited Aggregation Applet

Wed, 09 Jun 2010 14:05:00 +0000

Yes, I wrote an applet. I know it is very 1990s but, amazingly, it still does the job quite well. Ok, next time I should really use Flash to do this.

The Applet simulates Diffusion Limited Aggregation as described in Chaos And Fractals from Peitgen, Juergens, and Saupe. It represents ions randomly wandering around (in a Brownian motion) until they are caught by an attractive force in electrochemical deposition experiment. This kind of phenomenon occurs at all scales, for example it happens in the distribution of galaxies. You can play around with the applet at http://31416.appspot.com/dla.vm

Java & 3D Surface

Fri, 18 Dec 2009 16:07:00 +0000

I have been looking all around the web for a Java library that can draw a simple 3D surface. And I did not find any. Most charting library, including the well known JFreeChart, can only draw 2D charts.

I am quite shocked that something that has been in Excel for 15 years is still not available in Java. And it’s not easy to make your own.

double[][] Is Fine

Thu, 26 Nov 2009 14:51:00 +0000

In my previous post, I suggest that keeping a double[] performs better than keeping a double[][] if you do matrix multiplications and other operations.

This is actually not true. I benchmarked 3 libraries, Colt (uses double[]), Apache Commons Math (uses double[][]) and Jama (uses double[][] cleverly). At first it looks like Jama has a similar performance as Colt (they avoid [][] slow access by a clever algorithm). But once hotspot hits, the difference is crazy and Jama becomes the fastest (Far ahead).

JDK 1.6.0 Linux 1000x1000 matrix multiplication on Intel Q6600
loop index	Colt	Commons Math	Jama
1	11.880748	24.455125	9.828977
2	11.874975	24.265102	9.848916
3	9.772616	14.374153	9.826572
4	9.759679	14.368105	2.655915
5	9.799622	15.238928	2.649129
6	9.780556	14.741863	2.668104
7	9.72831	15.509909	2.646811
8	9.79838	15.724348	2.646069
9	9.726143	15.988762	2.646052
10	9.784505	15.121782	2.644572
	We don't include matrix construction time, and fetching the result. Only the multiplication is taken into account.

The difference is less pronounced on smaller matrices, but still there. Jama looks very good in this simple test case. In more real scenarios, the difference is not so obvious. For example Commons Math SVD is faster than Jama one.

Java Calendar Is Broken On JVM Upgrade

Wed, 19 Aug 2009 13:47:00 +0000

We ran into an interesting issue with TimeZone and Dates. If you print the same date on different JVMs, it might show a different printed date.

The reason behind this is the daylight saving time conventions. An old JVM won't necessarily have the same daylight saving time for a given TimeZone than a latest JVM, and therefore will interpret the date differently.

Here is the output of a very simple program on 2 different JVMs. The number is the long number used to represent the date. I only use SimpleDateFormat with different TimeZone:

JVM 1.5.0_12
loading date=Sat Jul 18 06:59:36 CEST 2009, number=1247893176505
using formatter in EST: 7/17/09 11:59 PM
using formatter in Asia/Singapore: 7/18/09 12:59 PM

JVM 1.5.0_20
loading date=Sat Jul 18 06:59:36 CEST 2009, number=1247893176505
using formatter in EST: 7/18/09 12:59 AM
using formatter in Asia/Singapore: 7/18/09 12:59 PM

The source code:


public class DateBug {       
   
    private static String FILE_NAME = "datebug.txt";
   
    public static void load() throws IOException {
        FileReader fr = new FileReader(FILE_NAME);
        BufferedReader br = new BufferedReader(fr);
        String l = br.readLine();
        br.close();
        long time = new Long(l);
        Date d = new Date(time);
        System.out.println("loading date="+d+", number="+d.getTime());
        SimpleDateFormat formatter = new SimpleDateFormat();
        formatter.setTimeZone(TimeZone.getTimeZone("EST"));
        System.out.println("using formatter in EST: "+formatter.format(d));
        formatter.setTimeZone(TimeZone.getTimeZone("Asia/Singapore"));
        System.out.println("using formatter in Asia/Singapore: "+formatter.format(d));
    }
   
    public static void saveNew() throws IOException {
        Calendar c = Calendar.getInstance(TimeZone.getTimeZone("EST"));
        c.set(2009, 06, 17, 23, 59);
        Date d = c.getTime();
        System.out.println("saving date="+d+", number="+d.getTime());
        SimpleDateFormat formatter = new SimpleDateFormat();
        formatter.setTimeZone(TimeZone.getTimeZone("EST"));
        System.out.println("using formatter in EST: "+formatter.format(d));
        formatter.setTimeZone(TimeZone.getTimeZone("Asia/Singapore"));
        System.out.println("using formatter in Asia/Singapore: "+formatter.format(d));
       
        FileWriter fw = new FileWriter(FILE_NAME);
        PrintWriter pw = new PrintWriter(fw);
        pw.println(d.getTime());
        pw.close();
    }
   
    public static void main(String[] args) throws IOException {
        System.out.println("JVM "+System.getProperty("java.version"));
        if (args.length == 1) {
            if (args[0].equals("save")) {
                saveNew();
            }
        } else {
            load();
        }
    }

What does this mean? This means that if you entered in a GUI in the first JVM a particular date & time using EST time zone. This will change when you read back in the second JVM.
This suggests that if you want to keep the same dates, you are better off saving in UTC where daylight saving time is not used and trick DateFormat. But I have to say this looks quite ugly.

Java int Overflow Behavior

Mon, 15 Jun 2009 16:15:00 +0000

A coworker recently asked me if there was a guaranteed behavior in the case of int overflow. He gave the specific example on:

can we rely that int x = Integer.MAX_VALUE + 1 is the same for every JVM on any platform?

I thought the answer would be easy to find in the Java specifications document. But I was wrong. It is not clearly defined.

I found a trick that suggests this behavior is indeed standard and will stay the same, it is related to type casting. Java guarantees that the cast of a long just truncates the long to int precision. Therefore if

long l = Integer.MAX_VALUE;
l = l +1;
x = (int)l;

Static Fields and Inheritance

Mon, 15 Jun 2009 16:02:00 +0000

Someone asked me recently to find out the real reason why the code from this thread fails. This is a fairly bad code, and not even a very good way to point out the problem. But the question is nonetheless interesting.

class Toto extends TotoParent{

 final static Toto a = new Toto ("a");

 public Toto(String a){
  super(a);
 }
}

import java.util.ArrayList;
import java.util.List;

public abstract class TotoParent {

 static List list = new ArrayList();
 
 public TotoParent(String a) {
  list.add(a);
 }

 protected static List get() {
  return list;
  
 }
}

import org.junit.Test;
import static org.junit.Assert.*;

public class TotoTest {

 @Test
 public void testGet(){
  assertEquals(1, Toto.get().size());
 }
}

I am quite used to static initialization, and would have answered the same as the first answer in the thread:
"Get is static and associated with TotoParent, so that is the same as calling TotoParent.get().size()". I would have even thought that the compiler would compile the call Toto.get() to TotoParent.get(). But running javap, you can see it is still compiled as TotoParent.get(). So there is still a lookup done. This is why the first answer is actually not that correct.

The important bit here is that Toto is never initialized, even if we call Toto.get(). The java specs (invaluable reference) explains clearly that calling a static method not declared in the class does not initialize the class.

Calling Toto.get() is not exactly the same as calling TotoParent.get().
If TotoParent.get() called another TotoSuperParent.get():
Toto.get() -> TotoParent.get() -> TotoSuperParent.get()
We compile then later we change to make TotoParent have a specific implementation of get(). Toto will then be automatically aware of it, without even recompiling it.

http://java.sun.com/docs/books/jls/third_edition/html/execution.html
paragraph 12.4.1

Benchmarking Languages Is Difficult

Wed, 03 Jun 2009 16:53:00 +0000

I often looked at the famous computer languages shootout for fun. Recently I noticed they had the infamous thread ring test. I posted not very long ago several blog entries about it showing how silly this test was.

Looking at the existing Java implementation for the test I decided to try to submit the tricky one using a pool of thread, and pooling message processing rather creating 1 thread per node. To my surprise, it was accepted without questions and I did have the best score for a Java program for a while. Shortly after someone else copied my program and got rid of various stuff not useful for the particular benchmark (breaking the interesting part of the design) and got accepted as well with of course a better result.

I decided to see if I could make an even more silly program - tailored for the test only. I managed to be orders of magnitude faster - 1 thread, no synchronization, everything processed in a FIFO (linkedlist) queue. This is actually a standard way to reimplement recursion. But I was honest enough not to hide that I consider that kind of program to cheat the test and got my entry in the "interesting alternatives".

In reality there is no difference in the "cheating" between my new program and the program that got accepted in the official list, they both cheat by using only 1 thread and process everything 1 by 1. There is not 1 thread per node in any of the program, and they can avoid any concurrency issues. One "looks" better because it uses a pool of 503 threads (but really use only 1 or 2 threads) and the other does not hide its use of 1 thread for processing. But this is not evident to people accepting the programs.

When I look at the haskell code, I can not really tell if it is creating 503 threads in the language or a pool or ..., you have to know each language quite well and sometimes it is not that easy to define what cheating is. Therefore this kind of benchmark is a bit disappointing. One should force the use of the same algorithm. But can you do so (a functional language won't use the same algo as a procedural one)?

Cholesky & Jakarta Commons Math

Fri, 15 May 2009 19:01:00 +0000

In Finance, Cholesky is a useful way to decompose Matrix. It is not so simple to find a BSD licensed code using cholesky (most of them are GPL like this one). There is one in Apache Commons Maths library, which is a very interesting library. However for performance, it is still not very practical for some things like Cholesky.

Looking at the source one can easily understand why. I did a small (many people will say not representative 1 million loop test) and finds out:

Java Logging Still Crap in 2009

Fri, 24 Apr 2009 15:01:00 +0000

When java logging API was first introduced in JDK 1.4 in 2002, it caused quite a lot a fuss around, with everybody asking “Why did not they just include Log4j instead of creating their own bastard child?”.

I remember having looked at it very shortly before continuing using Log4j on all projects I have been involved with.

Today, while doing a very small project, I tried once more to use java logging. The main reason is that I was lazy to add a dependency to one more jar for this small project. While trying I found out that:

Bachelier vs. Black

Mon, 23 Mar 2009 17:58:00 +0000

Black and Scholes gives a strange result for the price of a binary option under high volatility. You will learn here how to simulate a stock price evolution using Java, and how to show it using JFreeChart library. It starts with more complex concepts (don't be afraid) and goes done towards simpler things.

I could not write all that in a blog format, so I created a old HTML page about it here and a PDF version.

Senior Developers Team Productivity X4 (from MS Research Paper)

Tue, 10 Feb 2009 10:44:00 +0000

There is a very interesting MS Research paper about test driven development (TDD). It is one of the only real study about it that I know of. The paper conclusions from experiments over 4 TDD teams vs 4 traditional teams is:

"TDD seems to be applicable in various domains and can significantly reduce the defect density of developed software without significant productivity reduction of the development team"

Their data gives also other interesting results:

An experienced team (5 people over 10 years + 2 people under 5 years) : 155KLOC C# code (+60 test).

A junior team (3 people under 10 years + 6 people under 5 years): 41 KLOC Java code (+28 test).

If you do the ratio of KLOC/man month, you have the following graph:

I know this is very far from scientific evidence and more like astrology, but still, the most conservative ratio for senior/junior is 4.23!

The End Of Rings Around Plain Java - A Better Concurrency Test

Thu, 15 Jan 2009 15:54:00 +0000

In my previous post, I was wondering why single thread was faster. D Andreou gave the correct explanation: as we send only 1 start message and as each node only send 1 message to the next one, there is always only 1 message being processed. So the test is optimum on 1 thread. It does not make much sense to make a multithreading benchmark on a problem that is fundamentally single threaded.

Running Rings Around Plain Java - The Killer Code

Thu, 08 Jan 2009 13:21:00 +0000

I wrote my previous post too fast. I found a very simple change that increases the speed x6!

The idea is too process messages in a ThreadPoolExecutor. As my Nodes are Runnable, I just needed to initialize a common ThreadPoolExecutor, and in a sendMessage, execute the runnable each time.

Here is the full code: public class OptimizedRing {

    private ExecutorService executor;

    public static void main(String[] args) throws Exception {
        OptimizedRing ring = new OptimizedRing();
        RingNode node = ring.startRing(Integer.parseInt(args[0]));
        node.sendMessage(new StartMessage());
    }

    public RingNode startRing(int n) {
        RingNode[] nodes = spawnNodes(n, startTimer());
        connectNodes(n, nodes);
        return nodes[0];
    }

    private Timer startTimer() {
        Timer timer = new Timer();
        new Thread(timer).start();
        return timer;
    }

    private RingNode[] spawnNodes(int n, final Timer timer) {
        System.out.println("constructing nodes");
        long start = System.currentTimeMillis();
        executor = Executors.newFixedThreadPool(4);
        RingNode[] nodes = new RingNode[n+1];
        for (int i = 0; i < n ; i++) {
            nodes[i] = new RingNode(i, timer, null);
        }
        long end = System.currentTimeMillis();
        System.out.println("Took "+(end-start)+"ms to construct "+n+" nodes");
        return nodes;
    }

    private void connectNodes(int n, RingNode[] nodes) {
        System.out.println("connecting nodes");
        nodes[n] = nodes[0];
        for (int i=0; i<n; i++) {
            nodes[i].connect(nodes[i+1]);
        }
    }

    interface Message {
        String getType();
    }

    private static class StartMessage implements Message {
        public String getType() {
            return "START";
        }
    }

    private static class StopMessage implements Message {
        public String getType() {
            return "STOP";
        }
    }

    private static class CancelMessage implements Message {
        public String getType() {
            return "CANCEL";
        }
    }

    private static class TokenMessage implements Message {
        private int nodeId;
        private int value;

        public TokenMessage(int nodeId, int value) {
            this.nodeId = nodeId;
            this.value = value;
        }

        public String getType() {
            return "TOKEN";
        }
    }

    private class RingNode implements Runnable {
        private int nodeId;
        private Timer timer;
        private RingNode nextNode;
        private BlockingQueue<Message> queue = new LinkedBlockingQueue<Message>();
        private boolean isActive = false;

        public RingNode(int id, Timer timer, RingNode nextNode) {
            nodeId = id;
            this.timer = timer;
            this.nextNode = nextNode;
        }

        public void connect(RingNode node) {
            nextNode = node;
            isActive = true;
        }

        public void sendMessage(Message m) {
            queue.add(m);
            executor.execute(this);
        }

        public void run() {
            if (isActive) {
                try {
                    Message m = queue.take();
                    if (m instanceof StartMessage) {
                        log("Starting messages");
                        timer.sendMessage(m);
                        nextNode.sendMessage(new TokenMessage(nodeId, 0));
                    } else if (m instanceof StopMessage) {
                        log("Stopping");
                        nextNode.sendMessage(m);
                        isActive = false;
                        //
                    } else if (m instanceof TokenMessage) {
                        if (((TokenMessage)m).nodeId == nodeId) {
                            int nextValue = ((TokenMessage)m).value + 1;
                            if (nextValue % 10000 == 0) {
                                log("Around ring "+nextValue+" times");
                            }
                            if (nextValue == 1000000) {
                                timer.sendMessage(new StopMessage());
                                timer.sendMessage(new CancelMessage());
                                nextNode.sendMessage(new StopMessage());
                                isActive = false;
                            } else {
                                nextNode.sendMessage(new TokenMessage(nodeId, nextValue));
                            }

                        } else {
                            nextNode.sendMessage(m);
                        }
                    }

                } catch (InterruptedException ie) {
                    ie.printStackTrace();
                }
            }
        }

        public void log(String s) {
            System.out.println(System.currentTimeMillis()+" "+nodeId+": "+s);
        }

    }

    private static class Timer implements Runnable {
        private BlockingQueue<Message> queue = new LinkedBlockingQueue<Message>();
        private boolean timing = false;
        private long startTime;

        public void sendMessage(Message m) {
            //we don't need to change this implementation as timer is rarely called
            queue.add(m);
        }

        public void run() {
            while (true) {
                Message m;
                try {
                    m = queue.take();
                    if (m instanceof StartMessage) {
                        startTime = System.currentTimeMillis();
                        timing = true;
                    } else if (m instanceof StopMessage) {
                        long end = System.currentTimeMillis();
                        System.out.println("Start="+startTime+" Stop="+end+" Elapsed="+(end-startTime));
                        timing = false;
                    } else if (m instanceof CancelMessage) {
                        break;
                    }
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
            }
        }
    }
}

Code	Spawn	Send 100M messages
Scala Actors	15ms	270104ms
SimpleRing	11ms	493073ms
OptimizedRing (4 threads)	6ms	84727ms
OptimizedRing (5+ threads)	5ms	62593ms
OptimizedRing (1 thread)	5ms	60660ms

I finally saw my 4 cores used! Max multithreaded throughput is achieved at 5 threads. However 1 thread is faster. Is this related to memory bandwith limit?

Now I am left wondering if actors are really that important if one can achieve much higher throughput using plain Java and very simple concepts (BlockingQueue, ThreadPoolExecutor). Worse, this test is actually faster with only 1 thread...

Running rings around plain Java

Thu, 08 Jan 2009 12:26:00 +0000

Alex Miller has a very interesting test of Actors. He finds out Scala performance is relatively low compared to Erlang, and Kilim is very near Erlang. But Kilim code is the most difficult to read in the lot.

I thought it would be simple to just do the same test in plain Java. I wrote the code for it duplicating the scala logic using Threads instead of Actors.

public class SimpleRing {
    public static void main(String[] args) throws Exception {
        SimpleRing ring = new SimpleRing();
        RingNode node = ring.startRing(Integer.parseInt(args[0]));
        node.sendMessage(new StartMessage());
    }

    public RingNode startRing(int n) {
        RingNode[] nodes = spawnNodes(n, startTimer());
        connectNodes(n, nodes);
        return nodes[0];
    }

    private Timer startTimer() {
        Timer timer = new Timer();
        new Thread(timer).start();
        return timer;
    }


    private RingNode[] spawnNodes(int n, final Timer timer) {
        System.out.println("constructing nodes");
        long start = System.currentTimeMillis();
        RingNode[] nodes = new RingNode[n+1];

        for (int i = 0; i < n ; i++) {
            nodes[i] = new RingNode(i, timer, null);
            new Thread(nodes[i]).start(); //later use pool
        }

        long end = System.currentTimeMillis();
        System.out.println("Took "+(end-start)+"ms to construct "+n+" nodes");

        return nodes;
    }


    private void connectNodes(int n, RingNode[] nodes) {
        System.out.println("connecting nodes");
        nodes[n] = nodes[0];

        for (int i=0; i<n; i++) {
            nodes[i].connect(nodes[i+1]);
        }
    }


    interface Message {
        String getType();
    }


    private static class StartMessage implements Message {
        public String getType() {
            return "START";
        }
    }

    private static class StopMessage implements Message {
        public String getType() {
            return "STOP";            
        }
    }


    private static class CancelMessage implements Message {
        public String getType() {
            return "CANCEL";
        }
    }


    private static class TokenMessage implements Message {
        private int nodeId;
        private int value;

        public TokenMessage(int nodeId, int value) {
            this.nodeId = nodeId;
            this.value = value;
        }

        public String getType() {
            return "TOKEN";
        }
    }


    private static class RingNode implements Runnable {
        private int nodeId;
        private Timer timer;
        private RingNode nextNode;
        private BlockingQueue<Message> queue = new LinkedBlockingQueue<Message>();

        public RingNode(int id, Timer timer, RingNode nextNode) {
            nodeId = id;
            this.timer = timer;
            this.nextNode = nextNode;                        
        }


        public void connect(RingNode node) {
            nextNode = node;
        }


        public void sendMessage(Message m) {
            queue.add(m);
        }


        public void run() {
            while (true) {
                try {
                    Message m = queue.take();
                    if (m instanceof StartMessage) {
                        log("Starting messages");
                        timer.sendMessage(m);
                        nextNode.sendMessage(new TokenMessage(nodeId, 0));
                    } else if (m instanceof StopMessage) {
                        log("Stopping");
                        nextNode.sendMessage(m);
                        break;
                    } else if (m instanceof TokenMessage) {
                        if (((TokenMessage)m).nodeId == nodeId) {
                            int nextValue = ((TokenMessage)m).value + 1;
                            if (nextValue % 10000 == 0) {
                                log("Around ring "+nextValue+" times");
                            }
                            if (nextValue == 1000000) {
                                timer.sendMessage(new StopMessage());
                                timer.sendMessage(new CancelMessage());
                                nextNode.sendMessage(new StopMessage());
                                break;
                            } else {
                                nextNode.sendMessage(new TokenMessage(nodeId, nextValue));
                            }
                        } else {
                            nextNode.sendMessage(m);
                        }
                    }

                } catch (InterruptedException ie) {
                    ie.printStackTrace();
                }
            }
        }


        public void log(String s) {
            System.out.println(System.currentTimeMillis()+" "+nodeId+": "+s);
        }
    }


    private static class Timer implements Runnable {
        private BlockingQueue<Message> queue = new LinkedBlockingQueue<Message>();
        private boolean timing = false;
        private long startTime;


        public void sendMessage(Message m) {
            queue.add(m);
        }


        public void run() {
            while (true) {
                Message m;
                try {
                    m = queue.take();
                    if (m instanceof StartMessage) {
                        startTime = System.currentTimeMillis();
                        timing = true;
                    } else if (m instanceof StopMessage) {
                        long end = System.currentTimeMillis();
                        System.out.println("Start="+startTime+" Stop="+end+" Elapsed="+(end-startTime));
                        timing = false;                                        
                    } else if (m instanceof CancelMessage) {
                        break;
                    }
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
            }
        }
    }
}

I was a bit surprised by the result. It was slow and only 1 thread was really active at one time. This is why the test is particularly good. It is not trivial to reproduce the functionality in plain Java in an effective manner. It really shows how the concept of Actors can be useful.

More Double Troubles

Tue, 06 Jan 2009 12:12:00 +0000

We saw in a previous entry how one has to be careful with Double.NaN. Today we will see how regular double can cause problems. By the way the NaN issue was not Java specific and this issue is also general in different programming languages.

A coworker was shocked that in Java (I was a bit surprised he saw that only today, but it is true it can be surprising that such a simple thing does not work as expected):

408.16 - 40.82 = 367.34000000000003

In C, this would lead to the same result. This is all due to the binary represention of double numbers. Using the formula 2^(exponent)*1.mantissa where mantissa is on 52 bits, we have

408.16 decomposition:

exponent = 256. Then 408.16/256 = 1.594375 = 1 + 0x9828F5C28F5C28F5C28F5C... * 2^-52
We round to 52 bits, the mantissa is 0x9828F5C28F5C3 = 2676827028518339.
As a decimal, the internal value is (2676827028518339/2^52+1) * 256 = 408.1600000000000250111042987555265426635742

40.82 decomposition:

exponent = 32. Then 40.82/32= 1.275625 = 1 + 0x468F5C28F5C28F5C...*2^-52
Rounded to 52 bits, the mantissa is 0x468F5C28F5C29 = 1241304647293993
As a decimal, the internal value is (1241304647293993/2^52 + 1)*32 = 40.8200000000000002842170943040400743484497

The difference in decimal becomes 367.34000000000002472... which becomes 367.34000000000003 when represented in binary (to convince yourself you can apply the same technique).

The Solution

One solution to this problem is to use java.math.BigDecimal which stores a number as 2 integers, one for the digits, one for the exponent power of 10 (and not 2).
The correct code would become:

BigDecimal value = BigDecimal.valueOf(408.16).subtract(BigDecimal.valueOf(40.82));

value would then be 367.34.

But BigDecimal has also many potential for bugs. For example, you should never use the constructor taking a double but always the one taking a String.

new BigDecimal(408.16) = 408.16000000000002501110429875552654266357421875

This is because of the binary representation of 408.16 as a double. 408.16 is only an approximation of 408.16!

Another trick with BigDecimal is not to use equals(Object) but compareTo(Object) because 408.160 is not equal to 408.16 using equals.

Why Could not They Make it Work With Double?

If you were too lazy to follow the steps of the explanation. There is a simpler explanation. Imagine the representation of a number in base 3 with 2 "digits". Let's imagine 1/3 is represented as 0.1 (this is a very simple number representation) 1/3+1/3+1/3 becomes 0.1+0.1+0.1 = 1.0 (in base 3) = 1.0 if we convert to base 10. Now in base 10, 1/3 can only be represented as 0.3, so 1/3+1/3+1/3 = 0.3+0.3+0.3 = 0.9 <> 1.0.
So BigDecimal is only interesting to handle ... decimals! In the enterprise world, this should be most apps. It is a bit sad it appeared so late in the JDK. It should really be a primitive type.

Java Programmers Are Idiots?

Wed, 17 Dec 2008 20:16:00 +0000

My brother just sent me a funny quote. I don’t know if it is true or not:

entwickeln Sie lieber überzeugende Lösungen anstatt viele Stunden mit Coding zu verbringen? Ist Ihnen die Produktivität Ihres Teams wichtig?

Mark Driver, VP Research von Gartner, kommentierte kürzlich

“Here’s a simple equation. In terms of mental fortitude… 1 Smalltalk developer = 2.5 C++ developers 1 C++ developer = 1.5 Java developers”.

You don’t need german to understand. Of course it can not be true. How can anyone measure mental fortitude? And how does it related with productivity is another issue.

Double.NaN Is Evil

Tue, 09 Dec 2008 16:23:00 +0000

I don't know what Sun had in mind when creating Double.NaN number. It is very inintuitive to use. I am sure every single developer out there fell in the trap of trying to find out if a double was NaN or not using:

Double.NaN == myDouble

This does not work (I don't know the real reason why), one has to use:

Double.isNaN(myDouble)

Not intuitive!

Grails Spring Union Not Surprising

Thu, 27 Nov 2008 17:18:00 +0000

Looking out at some old post. I found out I was not far from the truth in January 2008 when I stated:“In 2008 the Ruby On Rails mentality will continue to prevail. In the Java world, Grails is the most likely to benefit from it. (…) It could also be something based around Spring as their current MVC solution is not very good and very old fashioned."I don’t think I will be right with the provocative Java is dead. A post recently titled No Future In Java makes some good points about where the future of Java still is: the web applications. Grails is probably today the best contender in the Java world, far ahead from the others, and it leverages the Java developers. However I am not sure one can say that RIA is a fad, or that RIA will only be done in a super powerful browser in the future. Microsoft might be a game changer here. The real advantages of the browser application so far are: easy deployment (the killer argument IMHO), “simple” security. I would not be surprised if in the near future, Microsoft advertises a solution for RIA easily deployed, based on standard protocols (HTTP?), a bit like IBM does with Eclipse RIA, but much more ambitious.

Java Is Dead

Fri, 21 Nov 2008 20:24:00 +0000

These days, I have the feeling that Java is dead. Even if, or maybe because I have used Java so much in the past 10 years, I have this feeling.

In 1998 Java was revolutionary. It was a very simple to learn object oriented language with modern concepts and familiar syntax. Furthermore the standard library had neat features like internet networking and it could be integrated in the browser. All this at a time when the internet just started to be popular.

OO Desillusion

Thu, 28 Aug 2008 17:00:00 +0000

I have read many inspiring books about object oriented programming. I find B. Meyer Object Oriented Software Construction one of the best in the lot. B. Meyer tries to explain in a progressive way why OO is better, by introducing it bit by bit. I have read different related design patterns book, the GoF one, Martin Fowler ones. I have been programming Java for about 10 years now. And yet, today, I feel unconvinced.

Java Compilation Performance / Processor Speed

Tue, 01 Jul 2008 19:50:00 +0000

I just found out my laptop was faster in default settings than my home desktop to compile a resonably sized project (5 min vs 6 min). I was surprised as I thought the disk in the desktop would make a big difference. The processor in my desktop is not that great (simple pentium e2180). My laptop has a 2ghz core2duo processor. In Ghz processor are of the same speed.

I tried to overclock my home pc to see what difference it could make, I made it run at 2.6Ghz instead of the standard 2Ghz. The compilation time dropped to 4.5 min.

DecimalFormat Is Broken

Wed, 14 May 2008 18:19:00 +0000

A friend of mine recently noticed that the good old DecimalFormat class is "broken". If you try to parse a string that is not a number but is starting with a number, the DecimalFormat.parse will return what it managed to parse.

The correct behavior should be to throw a parse exception IMHO. Judging from an old post in the Sun bug tracker, The folks at Sun don't think it really is, they call the default mode of parsing the "lenient" mode. It accepts bad inputs. Then why throwing ParseException at all and why not return 0/NaN when the first character is not a number? Why accepting 1toto2 as a number and not toto2?

In reality it can really create unexpected problems. For example, in France, 0.1 is 0,1 because of the Locale conventions. If a user enters 0.1 in a French Locale, a method using DecimalFormat.parse will interpret it as 0 without throwing any exception.

Note that DateFormat does not have that problem, at one point Sun added setLenient flag to be able to be in non Lenient mode. It would be very simple to do it with DecimalFormat, I did it myself as an exercise. In DecimalFormat.subparse, the 2 last break statements should stop processing in lenient mode. Lines 1528 to 1531:
sawExponent = true;
}
break; // Whether we fail or succeed, we exit this loop
}
else {
break;
}
become:

sawExponent = true;
} else {
if (isLenient()) {
parsePosition.index = oldStart;
parsePosition.index = oldStart;
return false;
}
}
break; // we succeed, we exit this loop

}
else {
if (isLenient()) {
parsePosition.index = oldStart;
parsePosition.errorIndex = oldStart;
return false;
}
break;
}

Using MiG Layout For Better Swing Development

Tue, 29 Apr 2008 18:27:00 +0000

I have forgotten a few libraries in my Better Swing Development article, and notably MiGLayout.

GridBagLayout is too verbose, and still feels too clumsy. This is why a while back I wrote a small tool to help visualize various GridBagLayouts for people who are not used to it. But it would have been much simpler to use a better layout instead.

MiGLayout is good, I managed to have good results without almost any practices on not so simple layouts. It also makes the code more concise.

SwiXml Review

Thu, 24 Apr 2008 12:19:00 +0000

Externalizing the screen layout is a natural step for many applications. It can be because a customer might want a slightly different presentation, or just for the sake of clearly separating layout code, or for using a kind of screen generator. Back in 2003, I wrote a small utility to help with GridBagLayout, nothing fantastic, but there was the feature of saving a generated layout in XML (or Beanshell) and using it in an application.

SwiXml saves a bit more than a layout as it instantiates the components as well, allowing to represent a hierarchy of components in XML. But it does not more than that. The "no generation" approach is interesting, I bet the author thinks the XML should be kept concise enough to be used easily without using a kind of generator. the XML is almost a mirror of Java code, but fortunately makes layouts more compact and easier to read. I tried version 1.51, and while it is quite well done, its documentation is a bit lacking and to me an important functionality is missing:

There is no "component" tag by default. One needs to explicitely name the class of component we want to add. I think there could be a component tag that instantiate an 0 parameter constructor by default, using reflection for example. The class of component to instantiate would be found through the field name.

Fortunately SwiXml is quite well done and this is not that big of an issue since you can just interpret new tags with 1 line of code:

engine.getTaglib().registerTag("datecombobox", DateComboBox.class);
for a Jide DateComboBox.

I said documentation was lacking, because in my first non Hello World trial, I wanted to use the standard constant names for GridBagConstraints (RELATIVE,REMAINDER, FIRST_LINE, LAST_LINE, ...). Putting <gridbagconstraint anchor="FIRST_LINE">did not work. By reading the code, I found out you could achieve it but you have to do <gridbagconstraint anchor="GridBagConstraint.FIRST_LINE">.

In the end SwiXml left me with a good impression. It is non intrusive, and handle well enough what it is done for. One major drawback, you can only use either BorderLayout or GridBagLayout or FormLayout (from JGoodies).

Better Java Swing Development

Thu, 17 Apr 2008 20:55:00 +0000

It has been a while since I have looked at the different options for building GUIs in Java. I like the approach taken by Eclipse, they provide not only GUI components, but a complete GUI application framework. Several years ago, I have stolen ideas from it (how they manage plugins) and applied it for Swing applications. This worked out quite well. The Netbeans platform now offers a more interesting alternative if you have to do Swing. The major drawback when compared to Eclipse, is that since the beginning Eclipse was all about making a platform, while Netbeans copied that idea only relatively recently and is less active about making it so. IBM uses Eclipse foundations for many of their products successfully.

Unfortunately on many projects, a Swing application is already there and all you have to do is add more screens, more forms. What I am trying to see here, is what could help Swing development is those cases.

JGoodies

JGoodies is old now, but can be quite useful in creating nice looking forms quickly. It is basically all around a layout better tailored to forms that the can do everything GridBagLayout.
It can help but it does not make a significant difference vs good practices with the GridBagLayout.

Jide Software

Jidesoft provides much more than JGoodies, but is not free (except the Jide Common Layer). I have used it in the past and found their API quite clean and easy to use. They offer more elaborated Swing components that should be part of the JDK. They started to go to the platform area only recently with their Jide Desktop Application Framework, which is not free either. I am not sure I would base an application around an unfree framework.

SwiXAT

It is a recent BSD licensed project based on top of Swix. It is basically MVC for Swix, with command binding in XML. I need to look more into it, it seems to provide an "application framework" as well where you have to launch your app through a SwixAt class (which IMHO sucks). I hope it is not a requirement and that only parts of it can be used and integrated into an existing application. I will hopefully have time to try it more concretely and will blog about it. In any cases it looks interesting.

SwiXml

An Apache licensed project, it seems like a decently used way of laying out swing components in XML, inspired by XUL. It might just be enough. Or is it just too similar to Java code translated to XML? I need to investigate this more.

JavaFX

JavaFX is probably the serious alternative to SwiXml. I like the idea of a better grammar to lay out and assemble components instead of using XML. Unfortunately it seems a bit invasive to start using for a part only of an application

I took a quick look at Swing Application Framework (JDNC2 redone once again). While it has interesting use of annotations, it seems very far from what Eclipse offers. And I don't feel like it will improve working with Swing a lot. I think except JavaFx, Sun seriously fucked around during many years with Swing. IBM has been much more focused and effective.

The 2008 Java Web Framework

Tue, 22 Jan 2008 11:50:00 +0000

One of my first web framework was a custom one, based on the book Web Development with Java Server Pages. That was in 1999, when Struts was not yet known. But it was web-MVC, very near from what Struts gave later, without the custom tag stuff (custom tags were not yet part of JSP specifications).

Until 2007, Struts was IMHO the most used web framework and the default choice, even though many contenders appeared, with more features, better code. 2007 saw the maturity of some of these framework. As similar framework to Struts, I particularly like Stripes, because the API is very lean. Some component frameworks got a lot of buzz as well, Tapestry some time ago, and today Wicket. I tried both, and while they are very interesting, and can be very productive for some tasks (web application more than web site), I am not convinced they are the tech of 2008. Web Applications are better done with Flex, or GWT.

The major news in the past few years has been Ruby On Rails. Everybody tried to replicate it in every language (with their own twist): Django in Python, Lift in Scala, ErlyWeb in Erlang, Cake in PHP, Grails (along with now abandonned Sails, and less populair Trails) in Java. And some people fought for making the point they had the idea before Rails. But Rails is what made Rails-like development popular.
After trying Rails soon after it was out (way before 1.0) and more recently (just before 2.0), I was amazed by the progress of Rails. I was impressed in 2004 about Rails, because the approach was new and effective. I disliked a bit the choice of Ruby. In the end in the project we ended up rewriting the persistence layer because our application was very specific. And I ported the app very quickly to Java. In 2007, I was even more impressed, and did not need to dwelve into arcanes of Rails anymore, but merely using it was enough to do lots and lots of things.

In 2008 the Ruby On Rails mentality will continue to prevail. In the Java world, Grails is the most likely to benefit from it. But I would not be surprised if there was another new Rails-like contender in the game (Rails on JRuby?). It could also be something based around Spring as their current MVC solution is not very good and very old fashioned.

Apache DbUtils Completely Useless

Fri, 02 Nov 2007 17:17:00 +0000

I am disappointed about the Jarkarta Commons DbUtils project. I give a link to it, because it’s a bad project (even if written in clean code). It is very simple, but it really does not do much for you.

I was looking for a very simple abstraction of JDBC. I thought bringing Spring in my project would be overkill. After trying DbUtils, I think again. It does not help. It does not handle frequent cases well, and it does not save many lines of code.

I fell in the trap of Boolean.getBoolean()

Fri, 12 Oct 2007 16:10:00 +0000

I was struggling to find a bug in a very simple application, it ended up being something as simple as using the damned Boolean.getBoolean("true") call instead of Boolean.valueOf("true").booleanValue() call.

The Boolean.getBoolean method is something you almost never need to use, as it checks if a particular system property is true or false. There is a similar method for Integer.getInteger, and a quick google search shows I am not the only one to think those method should never have been part of the basic API for Boolean/Integer. It is too easy to confuse with parseBoolean/parseInt, especially as parseBoolean does not exist in JDKs prior to JDK 1.5 (parseInt is older).

I can not imagine the improductivity this method has produced given its part of one of the most used class in the world.

Fast Web Development With Scala

Tue, 25 Sep 2007 14:02:00 +0000

I am currently experimenting with Scala. It seems quite convenient for web applications. Using Tomcat, it is possible to have a very productive developer environment.
Here is a sample Embedded Tomcat you can start in a Scala project:

import java.io._;
import org.apache.catalina._;
import org.apache.catalina.startup._;

object TomcatScalaServer {
  
  val CATALINAHOME : File = new File("../newsbeef.com");
  val WEBAPPS : File = new File(CATALINAHOME,"webapps");
  val ROOT : File = new File(CATALINAHOME,"web");
  val HOSTNAME : String = "localhost";
  val PORT : int = 8080;
  
  def await() {
     while( true ) {
          try {
            System.out.println("sleeping 100s");
              Thread.sleep( 100000 );
          } catch {
            case ie : InterruptedException =>;
          }
      }
  }
  
  def start() {
    val server = new Embedded();
    server.setCatalinaHome(CATALINAHOME.getAbsolutePath());

    val engine = server.createEngine();
    engine.setDefaultHost(HOSTNAME);

    val host = server.createHost(HOSTNAME, WEBAPPS.getAbsolutePath());
    engine.addChild(host);

    val context = server.createContext("", ROOT.getAbsolutePath());
     context.setParentClassLoader(Thread.currentThread().getContextClassLoader());
     context.setReloadable(true);
    host.addChild(context);

    server.addEngine(engine);

    val http = server.createConnector(HOSTNAME, PORT, false);
    server.addConnector(http);

    server.start();
  }
  
  def main(args: Array[String]) {
    start();
    await();
  }

}

Here is a sample Scala Servlet outputing html directly. This is a simple example, but it shows something important. With Scala, the view layer can just be regular scala classes. There is no need for JSP or other templating languages as Scala already embbeds XML very nicely. By using the reloadable feature of Tomcat (there are also other pure Scala ways) and Eclipse autocompile, changes are instantanously taken in account.

import javax.servlet.http._;

class ScalaServlet extends HttpServlet {

  override def init() {
  }
  
  override def doGet(request : HttpServletRequest , response : HttpServletResponse) = {
    service(request, response)
  }
  
  override def service(req: HttpServletRequest,resp: HttpServletResponse) { 
    val pw = resp.getWriter();
    var output = <html>
    <head><title>Scala Servlet Test</title></head>
    <body>
      <h1>Hello World!</h1>
    </body>
    </html>
    pw.println(output);
    pw.flush();
  }
}

Now I am eagerly waiting for improvements in the Eclipse Scala plugin (Organise imports, class navigation).

Spring Web Services, Finally!

Thu, 23 Aug 2007 11:22:00 +0000

Spring Web Services seems to be the technology I have been looking for recently. I am not a Spring bigot (too XML oriented), but here the Spring folks have something right.

I used to work with Web Services the simple way: create a java class (or EJB), expose it as Web Service through Axis or RAD, generating the WSDL in the process. And then a client would just be the reverse, take the WSDL, use a tool (Axis or RAD) that creates client Java classes from it automatically. Simple, easy.

Original Pattern: ServletRequest in ThreadLocal

Thu, 02 Aug 2007 12:07:00 +0000

After seeing Scala had elements of Erlang through Actors, I decided to take a closer look at the language. There is an interesting new web framework in Scala, called Lift. One drawback of Lift is that it seems to be very cutting edge and not that easy to grasp. While reading its source code, I stumbled upon a strange pattern: Storing the ServletRequest in a ThreadLocal.

I had not seen that before, and was wondering why one would do such a thing. It seems to be unintuitive. I found my answer through… GWT widgets. In this page, the author explain motivations behind doing such a thing:

Why Eclipse Is Better

Fri, 27 Jul 2007 14:37:00 +0000

Initially I adopted Eclipse instead of Emacs because it was more powerful to search code, and it allowed refactoring. I regularly tried other IDEs but always end up back to Eclipse, even though there has been less big improvements in Eclipse in the past years (but lots of small ones).

I just saw today that Eclipse allowed programmatic refactoring. Now that’s something quite amazing, and I don’t think other IDEs do that yet. Someone even had fun writing an Eclipse extension in Scala to add a particular kind of refactoring to Eclipse.

Tapestry5 vs Wicket: 1 - 0

Wed, 11 Jul 2007 12:49:00 +0000

Getting started with Tapestry 5 is easier than with Wicket 1.3. Some readers will complain that it is again the view of someone who has no deep knowledge of either Tapestry or Wicket. But I think it is important for projects to be easily accessible to developers. Wicket seems to have more buzz around these days, and has a detailed wiki with plenty of useful information in it. But that's the problem I see with Wicket, it is not simple to do simple things, that is why there is so much information to do simple things in the Wicket wiki.

Granted my test was based on a specific case for component frameworks, I was not so much interested into statefulness, I wanted to display a bookmarkable "user page" with content coming from hibernate.This kind of behaviour is quite general in web applications, especially in web 2.0.

It was relatively easy to have the page working with Wicket, although I was disappointed at their hibernate integration. Hibernate integration in wicket means either using the full databinder project, or creating your own solution. I chose the later based on source code from databinder, but I actually rewrote everything in the end. I was disappointed that databinder, a specific Hibernate oriented framework did not really handle Hibernate sessions the simplest way possible. Tapestry5 got that right. To manage Hibernate sessions right, I had to dwelve into Wicket code as no documentation offers insight about inner workings of wicket. The code was too complex for my taste. In my short experience, I saw it seemed the developers are changing it to the better, removing some unnecessary abstractions.

In the end I got frustrated many times with Wicket, and did not manage to have a bookmarkable page the way I wanted. You can have a bookmarkable page, but after some action on the page, it would become unbookmarkable. Furthermore, the structure of the URL is not very flexible without yourself rewriting completely the bookmarkable page feature of Wicket.

With Tapestry5, I was at first worried about the small amount of documentation on the site, the use of maven in the tutorial. I was wrong, documentation proved to be exactly what I needed, and detailed enough. It is much easier to understand how Tapestry5 works after reading the doc than Wicket. Concepts in Tapestry5 are simpler and more powerful. Maven use is in the end not that big of a deal, I am still not as comfortable with it but I am productive enough that it is not an issue, much more productive than with Wicket. The standard tutorial setup is a very good one.

Doing a bookmarkable page was trivial, it also was easy to have the format i wanted, and it was kept after action in the location bar. Hibernate integration was trivial, since Tapestry5 provides the tapestry-hibernate module, a few classes that helps managing the session and transactions for you. The only drawback is maybe the yet another inversion control system to learn. Tapestry5 IoC is very near from Guice in its philosophy. I wish Guice was made the default for IoC in Tapestry5.

To conclude, there is no doubt about it, Tapestry5 is the winner.

NetBeans 6.0M10 out without announcement yet!

Sat, 30 Jun 2007 07:15:00 +0000

I just found it while browsing netbeans website, here is the link. Netbeans is starting to be much more interesting that it used to be before 5.5, even though shortcuts are a pain, because so different from most other editors, and not always defined for important tasks. I like the all integrated feeling without plugin and slugishness by default.

Use ORM For Better Performance

Tue, 12 Jun 2007 12:59:00 +0000

This is not something I would have though a few years ago. It is something I learnt after working on many different projects, some using an ORM layer like Hibernate, Entity EJBs, or JDO, some using JDBC approach via Spring Templates or custom frameworks. Many projects that use ORM have performance problems, that don't seem that common with projects using JDBC. But the size of the database model of ORM projects is often much bigger than the one of JDBC projects (which actually makes sense). If you have only a few queries to do, why bother with ORM? This would be complexity for nothing.

But for most enterprise projects, the size of the database model is quite big, and the model itself can be complex (many relations between many tables). With this kind of model, ORM is more efficient. It is faster to develop with, creates less bugs due to string misspelled, or types badly read. It is also better performing. Doing 1 giant query to retrieve everything in 1 step is not faster, especially if you don't always need all the information retrieved. In a complex model, many cases are specifics, only useful in 10% of the cases. The temptation is high with a JDBC approach to do one giant query, because it is substantially longer (and more work) to do N queries. With ORM, it is a bit the opposite, by default N queries is easier to do. The problem is that N(ORM) tends to be very high if one is not careful with the mapping to avoid the N+1 problem. However it is simpler to reduce the number of queries by joining tables, rather than splitting queries, ORM performance optimization feels more natural.

Martin Fowler tends to be also pro ORM in its "Domain Logic and SQL" article. He also mentions something interesting about SQL query optimization:

It's also worth pointing out that this example is one that plays to a database's strengths. Many queries don't have the strong elements of selection and aggregation that this one does, and won't show such a performance change. In addition multi-user scenarios often cause surprising changes to the way queries behave, so real profiling has to be done under a realistic multi-user load. You may find that locking issues outweigh anything you can get by faster individual queries.

People Using Spring, EJBs Don't Know Basic JDBC

Wed, 30 May 2007 12:03:00 +0000

I recently found a bug in software we are developing. I traced it and found the root was improper JDBC handling. The application is written using EJBs, Spring and plenty of other relatively complex technologies. I was surprised that developers who were able to use all those technologies had no understanding of basic JDBC.

They fetched all the data (including double, decimal numbers) from the database as String using rs.getString() !

Wizards Bad For Productivity

Wed, 16 May 2007 14:56:00 +0000

IBM RAD comes with many wizards, to create EJBs, to create Web Services, do struts mapping… They are quite well done, making EJB < 3.0 usable, and Web Services look simple.

But wizards sucks at:

typos correction
repetition

But when you do a typo in your wizards, then all the files generated/changed are wrong, and you don’t necessarily know if you can just do a search and replace. Plus you don’t necessarily know all the files that were affected by the typo.

Less Productive With Maven2.

Sun, 29 Apr 2007 17:10:00 +0000

My first trials of Maven were failures. As I am stubborn, I tried again, on a new project, a quite simple one. It works, but it makes some easy things overkill. And the default way of using it makes a developer lose lots of time.

If I have a project with common classes, a standalone app, and a web app, then logically you do 3 projects, 2 of them depending on the common one. That’s how the default maven setup works, and that’s what their documentation presents. Now when using maven eclipse, this will create 3 project, none depending on each other. If you modify something in the common code, it won’t be seen by any of the other code, you have to publish it with maven first, this takes way too much time. Furthermore I did not see any way to force rebuild the common automatically from one of the other project. If you modify code in common and web app project, you need to call maven twice. I find all this very counterproductive, because you do those steps extremely often. Now there are probably some ways to do that with Maven2, but it is not the default behavior. I could add project dependencies in eclipse manually, and forget about maven while working in eclipse, but then the maven eclipse plugin is really useless. And you’ll face the same issues when you want to use maven tomcat deploy.

Java Concurrency In Practice Book Review

Wed, 14 Feb 2007 16:33:00 +0000

My reference book on Java concurrency is Doug Lea's Concurrent Programming in Java: Design Principles and Patterns. He is one of the authors of this new book, Java Concurrency In Practice. There is also Joshua Bloch, author of Effective Java, that many people love (but I am less a fan of it, even if I would recommend it to Java newbies), and author of Java Puzzlers, that I found more fun.

With such authors, I had relatively high expectations. I was surprised that there is not much material in common between Doug Lea's book and this one, which is a good thing. It's a different presentation, that focuses on different problems.

There is some very interesting material, I especially enjoyed chapter 16 on the Java Memory Model. There are not many books with information on it. I did not know that initializing a final member in the constructor was providing thread safety for the accessor method of the corresponding field if there is no other modifying method, no synchronization needed. And not having the member declared as final was breaking the safety.

I also had not heard of Amdahl's law before. It seems quite intuitive, but then the example in the book about processing parallel tasks using a synchronized LinkedList (improvement up to 3 threads, no more later due to time spent on synchronization) shows that the reality is not that intuitive.

I enjoyed how the book is written, and how information is presented. There is for example their concept of publication and escape (in chapter 3) which tells you important things, like not starting a Thread in a constructor (and if you still want to do it they present a nice way to do it).

I think this book is a simpler read that Doug Lea original book, and probably a better introduction (that still goes very deep) to Java concurrency programming.

Why are you not using Entity EJB 1.1?

Wed, 07 Feb 2007 16:38:00 +0000

I am currently quite busy learning about Websphere Commerce. I was surprised to notice that they still use Entity EJBs 1.1. My experience was that many companies turned away from Entity EJBs, sometimes from the beginning, preferring TopLink or Hibernate (later).

I know that Entity EJBs are more heavy to use than Hibernate, but with the proper tooling support, it’s not really a big issue. Are there more fundamental reasons? There are probably some mapping limitations, but IBM implementation is apparently quite good in term of features.

Social (Web) Revelation: Tags Are Lame

Tue, 16 Jan 2007 10:26:00 +0000

I had a revelation yesterday while thinking about del.icio.us, the most well known social bookmarks site. I find out that I almost never use my tags, and I am often not satisfied by results when I search using tags. What use can you make of 100s of tags? in the end only less than 10 are usable to classify very different stuff. And even then compared to a search it’s not a much useful classification.

1 year of mac mini - the deception point

Mon, 08 Jan 2007 15:56:00 +0000

I am less enthusiastic about the Apple experience than I was one year ago.

I am quite enthusiastic with the hardware in general, except the keyboard. The computer is small and quiet, and there are not many competitors at this price on the market.

Experience as a developer

I feel frustrated with MacOs X as a development environment, when compared with Ubuntu Linux. I always miss the multiple desktop, despite expose magnificient look. The standard behaviour for the terminal, bash, vim, etc. are more convenient in Linux. Installing software and updates is much more convenient on Linux thanks to the open source repositories. The key bindings for special characters is awful for a programmer on the Apple keyboard, because non standard, and not even written on the keyboard. Installing linux on a mac mini intel is not trivial, but I am considering it seriously.

Tired Of Bad Singletons

Thu, 04 Jan 2007 11:50:00 +0000

While looking through some code for a project, I saw that:

public static final ProductYP instance = new ProductYP();

public ProductYP()
{
    if (instance != null)
      throw new RuntimeException("Only one instance allowed");

    prods = new HashMap();
}

public static ProductYP getInstance()
{
    return instance;
}

And I don't think it was done by a newbie... It's actually not far from being correct, it's just that the guy obviously does not know about private constructors. I have seen several broken singleton implementations in previous projects and had several debates on the double-checked locking pattern (it works since JDK1.5 with volatiles but is useless). I am upset to see another half broken implementation. Everybody should have read at least this.

What have you seen as broken singletons?

Declaring Your Logger - No Problem

Fri, 29 Dec 2006 11:45:00 +0000

I used to like the java specialists newsletter, most news used to be a bit challenging. Nowadays however the quality is lower. In the latest news, the author proposes to use the StackTrace to get the class name, in order to declare a Logger independently of any explicit reference to the class name.

While this is a clever hack, it still requires some code to be duplicated in every class, compared to Aspect/IoC approach. But was there really a problem with declaring loggers the usual way in the first place?

Very Quick Django Trial -> Bye Bye Django

Thu, 21 Dec 2006 17:48:00 +0000

I tried a bit django for one very simple thing, where django is not of much use, but it was a way to try it with something else than examples on the site.

Even if django templates features seem very good and very convenient when reading the docs (for example the extends keyword), I found out that the template system is very annoying. Firstly one has to learn another templating system. Secondly, and much more importantly, the debug facilities for templates are close to non existent. I have a better memory using ruby on rails templates (which are not the best either).

Java templates, based on Velocity, JSPs or JSF are much nicer to debug. PHP (which I consider an ugly language that can unfortunately often gets the job done) used to be crap at debugging, but with eclipse plugin phpeclipse it is much better these days.

I think templates play a very important role in web application development, and I am surprised that a framework with so many good ideas fails in reality (at least for me) because debugging templates is a pain. Hopefully it will become better one day, but until then bye bye Django.

A Proof That JavaScript Sucks.

Fri, 15 Dec 2006 16:51:00 +0000

Google developed GWT.

Why would GWT be so well acclaimed if JavaScript was a good language. When you talk about GWT to someone (a developer preferably), the first reaction is often

great I don’t need to do any JavaScript

I recently discovered another similar open source project, ZK. The first thing you can read on the website is

ZK is an open source Ajax Web framework that enables rich user interface for web applications with no JavaScript and little programming.

Building With Ant + Maven Antlib?

Thu, 14 Dec 2006 10:51:00 +0000

I discovered a new baby in the build world. Some open source projects use a combination of ant and maven. I thought you could vaguely use ant from maven. But I did not know of Maven Antlib that allows you to use maven2 from Ant. Why would you use Maven2 from Ant?

because Maven2 requires to adapt your project to its standard hierarchy, or to have excellent Maven2 skills to create your own archetype for each existing project.
because everybody knows Ant, but Ant does not manage dependencies well. Maven2 Antlib can be used just for that: managing dependencies.

The Hibernate Network Proxy, A Dumb Idea.

Mon, 04 Dec 2006 14:41:00 +0000

A friend of mine had a good idea for a Java project named “hibernate proxy”. I renamed it “hibernate network proxy” because hibernate proxies are already a well known concept among hibernate users (it allows lazy loading through a Java Proxy class).

The hibernate network proxy, is a way to use hibernate remotely, that is to proxy hibernate calls through TCP (or even HTTP) to another machine, that will execute the session.saveOrUpdate or the session.load, etc. calls and return the corresponding objects, serialize them and deserialize them for use on the client.

Workaround When Eclipse Fucks Off With CVS

Tue, 28 Nov 2006 11:42:00 +0000

Until recently, I always found Eclipse CVS support to be excellent. The synchronize view was really a big plus to do clean commits, without forgetting any file.

But I have encountered several problem with Eclipse 3.2+:

On one project, it keeps putting the build directory into CVS, and I can't find how to change that setting. Each time I open the project I have to change the CVS/Entries file.

On another, where I have lots of files in, Eclipse delete some of the files (checked in CVS) regularly, although not often. It won't retrieve them back again from CVS. It seems to be the same files each time. I have no clue why.

The diff sometimes tells me the whole file has changed while there is absolutely no change.

I could not make the merge work. It kept on complaining about some parameters.

And now the workaround for all those cases:

The good old CVS command line.

I never remember how do to a merge with the command line, so I each time go to google and read this merge example page. Here are the very simple steps if you are in a branch and want to merge:

cvs update -A mypoem.txt (use head as local repository)

cvs update -j release-1 (merge release-1 changes into local repository)

```
cvs update mypoem.txt
```
```
cvs commit mypoem.txt
```

Maven: Game Over, Try Again

Fri, 24 Nov 2006 17:51:00 +0000

I am writing only as a newbie on Maven2. I tried to use it several times on small projects, but each time I found out that the time involved to make it work with an existing project was too much a pain in the a**. When confronted to very simple hierarchies that are not the standard of the documentation, I was a bit lost on how to arrange things, except changing the whole hierarchy to adapt to Maven.

How JDepend Changed My Java Packaging

Fri, 13 Oct 2006 09:27:00 +0000

One important feature of Java language is the package keyword. It helps a lot in modularizing your code. But how exactly one should use it is not that clear.

I have been on projects where 100s of classes are in a same package and more often where you have packages for every 2 classes. Often, packages are chosen so as to split functionalities. But often as well, people are packages maniacs and create way too many of them, because they want to sort things out, not necessarily applying a consistent logic. For example, I am sure many of you have seen the “blahblah.exceptions” packages where you dump exceptions classes.

You Know IOException? Think Again!

Thu, 14 Sep 2006 10:59:00 +0000

I was amazed today to find out that there was no constructor IOException(String, Throwable) or IOException(Throwable) in JDK1.4 and JDK1.5. It is finally in JDK1.6, I can’t believe it took Sun that much time to change that.

So the workaround is:

`IOException ioe = new IOException("message");`
`ioe.initCause(e);`
`throw ioe;`

It can also be written as:

`throw (IOException) new IOException("message").initCause(e);`

It is not a major problem, but still. We can all thank the guy who reported that as a bug to Sun in 2004.

Good Software Books - 2006 version

Mon, 11 Sep 2006 14:27:00 +0000

Here is an update of the most interesting software books I found interesting. I already made such a list in 2005:

Object Oriented Software Construction, 2nd Ed, by Bertrand Meyer. This made me understand why OOP is important, what is important in OOP and why it is still relevant. It contains lots of important guidelines you can apply to better design programs. One famous quote is "Real systems have no top". Code Complete (Microsoft Press) covers some of the same ideas as Meyer's book, and some more pragmatic issues that arise in software projects ("measure twice, cut once"). I also liked the chapter on "table driven design".
Concurrent Programming in Java, Second Edition, by Doug Lea: there is all you need to know about programming in a multithreaded environment in it. The chapter on synchronization should be a must read for every Java developer.
Design Patterns by the GoF: simply the best presentation of most common design patterns. I have a glance on it once in a while.
Artificial Intelligence through Prolog, by Neil C. Rowe (Prentice-Hall): if you don't remember much about Prolog, it's a good book. It details how Prolog interpreters work.
File Systems Forensic Analysis, by Brian Carrier (Addison Wesley): everything you need to know about disks and file systems, every trick to recover lost data. To go more in depth into file systems theory, you can read the free Practical File System Design with the Be File System from D. Giampaolo, the creator of BeFS.
Mind Hacks (O'Reilly): you probably already bought that one. Not about software, but excellent.
Practical Issues in Database Management by Fabian Pascal (Addison Wesley): short but concise book on relational database theory. F Pascal is of CJ Date school of thoughts, "null are evil". You don't need to read it if you have read any other similar book (by CJ Date for example).
Inside The Java Virtual Machine, by Bill Venners (McGraw-Hill): there are other books on the same subject. While this book has some bad reviews, I found it an easy read, and it explain well enough for me all the inner details of Java. I found "Programming for the Java Virtual Machine" not better, and with some bad example of a Prolog language for the JVM (1 full chapter for this to present a way too simple thing to be of any use).
Programming Jabber, by O'Reilly: I have read it 2 years ago, I enjoyed how they made you go through building a Jabber server.
Lucene In Action: the only book about Lucene, some of the info is useful to understand how it is designed.
Hibernate In Action: if you have to use hibernate, this is the book to read. It presents different ways of using it.
some Javascript+DHTML(+CSS) book: useful if you have none. They often have good tricks to solve the usual problems. In the AJAX days, it becomes useful again. I find O'Reilly's "Javascript - The Definitive Guide" good to have an API reference (as ebook), and "Javascript and DHTML Cookbook" has the most useful recipes.
Lean Software Development by Poppendieck (Addison Wesley): easy to read, good presentation of software dev management problems and pragmatic solutions. My favorite on the subject. Another good one but very specific is "Requirements Management" (MS Press).
UML Distilled by M Fowler: always useful to have if you need to draw UML diagrams.

I Get Spring

Wed, 23 Aug 2006 18:04:00 +0000

When you google up Java Spring, one of the best results is a post from crazybob called “I Don’t Get Spring”. For a long time, I shared a similar opinion. But now that I have used it, I get it. I only use it for defining replaceable services, so when I talk about Spring, I mean spring-core and spring-beans.

These two packages are not big, and have only very few dependencies. So it is quite easy to use Spring in any project, be it small or not. It is actually easier than using NanoContainer because Spring has less external dependencies.

Java Serialization vs .NET Serialization - Java Perverse?

Tue, 22 Aug 2006 12:55:00 +0000

Did you know what happens in Java when you serialize a subclass of a non serializable class? I was surprised by the answer: it works!

Unfortunately it is not a good thing, because it will serialize fields from your subclass and no fields from the parent class. So you'll end up with a half serialized instance.

In .NET, it breaks at runtime, throwing an exception, which is I think, much more logical, because then you don't end up with half data somewhere.

Java Code:

import java.io.ByteArrayInputStream;
       import java.io.ByteArrayOutputStream;
       import java.io.ObjectInputStream;
       import java.io.ObjectOutputStream;
       import java.io.Serializable;
       
       
       
       public class Test
       {
         public static class Toto 
         {
           public String me;    
         }
         
         public static class Toto2 extends Toto implements Serializable
         {
           public String you;
           public String toString()
           {
             return me+" "+you;
           }
         }
         
         public static void main(String[] args) throws Exception
         {
           Toto2 t = new Toto2();
           t.me = "it's me";
           t.you = "it's you";
           System.out.println("t="+t);
           ByteArrayOutputStream b = new ByteArrayOutputStream();
           ObjectOutputStream oos = new ObjectOutputStream(b);
           oos.writeObject(t);
           ObjectInputStream ois = new ObjectInputStream(new ByteArrayInputStream(b.toByteArray()));
           System.out.println("u="+ois.readObject());
         }
       }

will output:
t=it's me it's you
u=null it's you

.NET Code:

using System;
using System.Collections.Generic;
using System.Text;
using System.IO;
using System.Xml;
using System.Xml.Serialization;
using System.Runtime.Serialization.Formatters.Binary;

namespace ConsoleApplication
{
    public class Toto
    {
        public string me;
        public override string ToString()
       {
            return me;
       }
    }

    [Serializable]
    public class Toto2 : Toto
    {
        public string you;
        public override string ToString()
       {
            return you + " " + me;
        }
    }

    class Program
    {
        static void Main(string[] args)
        {
            Toto2 t = new Toto2();
            t.me = "it's me";
            t.you = "it's you";
            using (FileStream fs = File.Create(@"c:\test.bin"))
           {
                BinaryFormatter bFormatter = new BinaryFormatter();
                bFormatter.Serialize(fs, t);
            }
            Console.WriteLine("t=" + t.ToString());
            Toto2u = null;
            using (FileStream fs = File.Open(@"c:\test.bin", FileMode.Open))
            {
                BinaryFormatter bFormatter = new BinaryFormatter();
                u = (Toto2)bFormatter.Deserialize(fs);
            }
            Console.WriteLine("u="+u.ToString());
            Console.ReadKey();
        }
    }
}

will throw an exception.

5 Minutes of Google Spreadsheets

Thu, 17 Aug 2006 12:10:00 +0000

Today I noticed a Google Spreadsheets link in my gmail screen. I had read about it but never bothered to try before. In the finance industry, a lot of traders use excel, so I was wondering if Google spreadsheets could be another fit.

Unfortunately for Google, under Linux at least, I don’t find Google Spreadsheets usable for anything else than storing and sharing some information, not often updated. Although I admire the engineers that managed to write the Javascript behind Google Spreasheet, it is way too slow for using it in interesting ways. Editing is slow, copy/paste is slow, sorting is slow.

Procedural Programming in an OO language

Tue, 08 Aug 2006 12:35:00 +0000

OO is an old buzzword, that is not required anymore to get an employment. Recruiter seems to prefer SOA, Web Services, and in France, “mutualisation” and “urbanisation”. Sometimes I really wonder if OO made it.

I am sure many of you are confronted with programmers that love procedural programming in many of your projects. They might use an OO language but in the end will organize everything by “type”, split invariably state and logic. Everything will be so much better stateless. And we will create lookup maps to glue everything back together. In a way I feel they remove the OO of the language.

Back To Real Java, Bye Bye J2EE

Wed, 26 Jul 2006 16:26:00 +0000

I changed job recently. In this new job, it is refreshing to see Java used like in the old days, without the J2EE layers, and without the extra IBM layers of my previous job. Granted, the fresh Java approach does not apply to many projects, because a lot of apps are just about interfacing a database with a web interface. But Spring success showed that even for many of those projects, fresh Java approach with a small framework is enough.

Moving weekly Javablog stats to a new blog

Mon, 26 Jun 2006 17:25:00 +0000

Javablogs.com top 10 weekly/monthly/yearly entries were starting to pollute my blog too much for my taste. It is more appropriate to dedicate a blog to them. That is what I should have done in the first place as it is very easy to have many blogs with blogger.com.

So you’ll find at javabuzz.blogspot.com the weekly top 10 most read entries on Javablogs, and sometimes more.

JSF Was Too Hard for Experienced Developers

Fri, 23 Jun 2006 14:59:00 +0000

While starting to look into Seam, I noticed that all examples use JSF for the view, and there is no alternative to it. If someone like Gavin (from Hibernate fame) thinks JSF is usable, then I probably overlooked something when I looked into it a few years ago, when JSF was the craze of the moment.

At the beginning, JSF looks very similar to ASP.NET. But I have a small experience with ASP.NET, and ASP.NET is quite simple to understand and use. You can throw up inexperienced developers at it, they will manage to create something that works quite quickly. ASP.NET feels quite natural once you understand the postback thingy. JSF is another beast. Maybe part of it is due to the fact that the Java world has no excellent tools integration like Microsoft has (IBM RAD, one of the most advanced concerning integration, is quite far off). But there are also many technological reasons; when I read that article about JSF shortcomings with JSP, I was shocked that even to do very simple things, you would screw up, because simple things can be very complicated in JSF.

On EJB Restrictions - Can You Open a Socket?

Mon, 19 Jun 2006 11:48:00 +0000

During job interviews, one of the frequent questions is

“What restrictions are imposed on EJBs?”.

I had a chat recently with someone about it, and we still don’t agree on the answer.

He believes we can’t open sockets, while I believe it is perfectly allowed, if you don’t retain the socket connection between method calls. His arguments are based on an analogy with files. After all in Linux, files are sockets. Here are his interpretations on a document about EJB restrictions on java.sun.com:

Non Blocking with Traditional Java IO - On the Use of InputStream.available() and Thread.sleep()

Thu, 08 Jun 2006 16:31:00 +0000

Some time ago I did quite a lot of IO in Java and I yet did not see this way of reading a InputStream from a Socket:

    InputStream in=channel.getInputStream();

    channel.connect();

    byte[] tmp=new byte[1024];
    while(true){
      while(in.available()>0){
        int i=in.read(tmp, 0, 1024);
        if(i<0)break;
        System.out.print(new String(tmp, 0, i));
      }
      if(channel.isClosed()){
        System.out.println("exit-status: "+channel.getExitStatus());
        break;
      }
      try{Thread.sleep(1000);}catch(Exception ee){}
    }
    channel.disconnect();

This comes from a piece of code from an example of JSch , a good ssh client in java. A work collegue had the bad idea to remove the Thread.sleep call and was struggling with why it would randomly work.The way I would have done it is the following:

    InputStream in=channel.getInputStream();

    channel.connect();

    byte[] tmp=new byte[1024];
    int bytesRead = 0;
    while((bytesRead = in.read(tmp,0,1024) >= 0){
      System.out.print(new String(tmp, 0, bytesRead));
    }

    if(channel.isClosed()){
      System.out.println("exit-status: "+channel.getExitStatus());
      break;
    }
    channel.disconnect();

This has the advantage of being more readable and having less secret spices in it. In the first code, the call to available() is non blocking, meaning that without the Thread.sleep(), there will never be the time for the socket buffer to fill up. But is the first code more efficient or the second code?

I did a search on google to understand the interest of the first code. The only advantages I found in the first code are the possibility to interrupt the thread running the code and a finer grained control on timeouts.

There is a lengthy explanation by Doug Lea in his book "Concurrent Programming in Java". This book usually provides excellent explanations, and is a must read for anybody doing concurrent programming. But this time, about this subject, I did not find him that clear.

There is a more simple explanation in a course from San Diego State University (see last example)

A read() on an inputstream or reader blocks. Once a thread calls read() it will not respond to interrupt() (or much else) until the read is completed. This is a problem when a read could take a long time: reading from a socket or the keyboard. If the input is not forth coming, the read() could block forever.

As usual, you should not rely on all what you read on the web, as this page (SCJP Questions & Answers) testifies:

Q. When will a Thread I/O blocked?

A:

When a thread executes a read() call on an InputStream, if no byte is available. The calling Thread blocks, in other words, stops executing until a byte is available or the Thread is interrupted.

Still I am wondering if the second code would not just go into IOException (socket timeout), on timeout (adjustable with Socket.setTimeout ) and release the Thread then. Do you have an idea when the first code could be better?

Is Java Flawed - a big advantage of Python/Ruby/(your favorite interpreted language)

Fri, 02 Jun 2006 17:33:00 +0000

Java is supposed to be much better to build big projects, because of static type checking, and all the rigour around the language. But how many of you have seen medium sized projects taking more than 30 minutes to build.

At work, they have a standard J2EE project, with only about 50 EJBs, hundreds of JDO classes, and standard classes. Between the JDO generation, EJB generation, EJB dependencies calculations, and packaging, it takes 20 minutes. And the project is not doing that much. One can optimize to avoid dependencies calculations and it would then take about 12 minutes. But still.

Using Linux to Recover Fucked Up Windows Data

Thu, 01 Jun 2006 01:52:00 +0000

Recently, one of my relatives computer under Windows XP, refused to boot. There was no way of fixing it with Windows Install CD as partition table seemed corrupt to Windows. I tried everything in an 2003 Ultimate Boot CD, but nothing worked out.

Someone gave me an install cd of Ubuntu Linux, and it managed to read the data. Well sometimes only. The erratic behaviour was due to a bad ATA cable. This probably was the cause of the corruption in the first place. Anyway with a new cable, Windows was still not able to read its data. But Ubuntu Linux, now working well, was able to, without having anything to configure (except mounting the drive).

Java HTML Parsing Example With htmlparser

Tue, 30 May 2006 11:02:00 +0000

Every week, I post javablogs top 10 most read blog entries on this blog. The reason for it was that I don't look at what's happening on the week-end and this will pickup interesting stories from the weekend, and I also don't watch javablogs everyday. Overall I find it quite good to be uptodate with interesting stuff happening on javablogs.

As mentionned in an earlier post my library of choice to do the parsing is htmlparser (on sourceforge) because it's free, open source and because I am lazy and did not want to do my own. If you know a better open source library, feel free to add a comment about it, I'll be glad to hear about it. htmlparser is not the easiest library to use, there are many entry points and it's not immediately clear which one to choose. So I post here how I used it if it can save a few minutes to people having to do this task.

  private static Entry parseEntry(String content) throws ParserException
  {
    final Entry entry = new Entry();

    final NodeVisitor linkVisitor = new NodeVisitor() {
      
      @Override
      public void visitTag(Tag tag) {
        String name = tag.getTagName();

        if ("a".equalsIgnoreCase(name))
            {
              String hrefValue = tag.getAttribute("href");
              if (hrefValue != null && !hrefValue.startsWith("http://"))
              {
                if (!hrefValue.startsWith("/")) hrefValue = "/"+hrefValue;
                hrefValue = "http://javablogs.com"+hrefValue;
                //System.out.println("test, value="+hrefValue);
              }
              if (hrefValue != null)
              {
                hrefValue = hrefValue.replaceAll("&", "&amp;");
                tag.setAttribute("href", hrefValue);                
              }
            }
      }
    
    };
    
    NodeVisitor visitor = new NodeVisitor() {

      @Override
      public void visitTag(Tag tag) {        
        String name = tag.getTagName();
            if ("span".equalsIgnoreCase(name) || "div".equalsIgnoreCase(name))
            {              
              String classValue = tag.getAttribute("class");
//                LOGGER.debug("visittag name="+name+" class="+classValue+"children="+tag.getChildren().toHtml());
              if ("blogentrydetails".equals(classValue))
              {
                Pattern countPattern = Pattern.compile("Reads:\\s*([0-9]*)");
                Matcher matcher = countPattern.matcher(tag.getChildren().toHtml());
                if (matcher.find())
                {
                  String countStr = matcher.group(1);
                  entry.count = new Integer(countStr).intValue();
                }
                
              }
              else if ("blogentrysummary".equals(classValue))
              {
                try
                {
                  tag.getChildren().visitAllNodesWith(linkVisitor);
                }
                catch (ParserException pe)
                {
                  LOGGER.error(pe,pe);
                }
                entry.description = tag.getChildren().toHtml();                 
                entry.description = entry.description.replaceAll("\\s+", " ");
              }
              else if ("blogentrytitle".equals(classValue))
              {
                try
                {
                  tag.getChildren().visitAllNodesWith(linkVisitor);
                }
                catch (ParserException pe)
                {
                  LOGGER.error(pe,pe);
                }
                entry.title =tag.getChildren().toHtml(); 
                entry.title = entry.title.replaceAll("\\s+", " ");
              }              
            }
            
      }

    };
    Parser parser = new Parser(new Lexer(new Page(content,"UTF-8")));
    parser.visitAllNodesWith(visitor);
        if (entry.title != null)
        {
          return entry;
        }
        else return null;
  }

First Steps With EhCache

Thu, 04 May 2006 18:10:00 +0000

If you need to cache objects in your system, Ehcache is a simple cache written in Java, widely used and well tested. I will present here a short tutorial on how to use EhCache for people who don't want to look around the documentation at first, but just want to test if it works in their project and to see how easy it is to setup.

Installation

Download Ehcache from the Download link on http://ehcache.sourceforge.net. Current release is 1.2.

Unpack Ehcache with an unpacker that knows the tgz format. For unix users, it is trivial, for windows users, 7zip is a free (and open-source) unpacker. It is probably the most popular, but there are other ones like tugzip or izarc or winrar.

In your java project you need to have ehcache-1.2.jar, commons-collections-2.1.1.jar and commons-logging-1.0.4.jar (versions numbers may vary) in your classpath, those libraries are shipped with ehcache.

Cache Configuration

Write an ehcache.xml file where you describe what cache you want to use. There can be several files per project, several cache descriptions per file. I use here a persistent cache. Configuration file is well described at http://ehcache.sourceforge.net/documentation/configuration.html

</ehcache>

Null vs. Errors

Wed, 26 Apr 2006 16:16:00 +0000

I am not particularly a fan of JCS (Jakarta Cache System) as I find ehcache code very clean and simple. But I have to say the author has some good comments on the site:

Nulls vs. Errors

I started to support ObjectNotFoundExceptions for failed gets but the overhead and cumbersome coding needed to surround a simple get method is ridiculous. Instead the JCS return null.

For having seen too many times the ObjectNotFoundException "pattern", I can only agree!

Algorithms in Java (Third Edition) Book Review

Wed, 26 Apr 2006 12:26:00 +0000

The book Algorithms in Java is huge, but unlike the usual huge books, the content is very interesting. It can be used as reference material, or as toilet book (to learn things while you're wasting time in the toilets).

You will learn simple things, like what is the "raison d'être" of linked lists. The author gives very good examples to illustrate his propositions. He explains through the sieve of erathostene and through Josephus problem the advantages of arrays or linked list.

You will learn step by step everything that is to be known in algorithms. Recursion, divide and conquer, Tree knowledge will be useful for the later sorting and searching chapters.

The chapter on Hashing will make you understand very clearly why the source of String.hashCode() is

public int hashCode() {

  int h = hash;

  if (h == 0) {

     int off = offset;

     char val[] = value;

     int len = count;

     for (int i = 0; i <  len; i++) h = 31*h + val[off ++];

     hash = h;

  return h;

There might be too much info on different types of sorting algorithms and the book becomes there more a reference book than anything else. But overall, you will learn plenty with this book. It is very well written, complete, and will refresh one's memory. I find it useful to read back things I learnt after a few years as I then have a very different view of the subject, and I pay closer attention to some details I completely missed the first time (sometimes).
Tags: java book review

Caching HTTP Responses in Java

Wed, 19 Apr 2006 18:14:00 +0000

Caching HTTP response can dramatically improve performance of your app if what you generate is in reality not very dynamic. There are many free caching frameworks in Java. Most popular seem to be ehcache, oscache, jcs and JBoss Cache.

ehcache is quite simple to use and its code is clean. They have a CachingFilter that you can put in your webapp server to cache transparently HTTP responses. However as the framework only allows you to store Objects (which makes sense for most uses), I was wondering how they cached the HttpResponse which is a stream. I was a bit disappointed by the answer, they just create a copy ByteArrayOutputStream and call toBytes() to store it in the cache. While this is optimal for a memory cache store (the whole response will anyway be in the cache, although I am not sure if they check for particularly big responses to avoid caching those or try to cache those) I don’t think it is that good for a disk cache store.

Javablogs.com 2005 Top 20

Tue, 18 Apr 2006 16:42:00 +0000

How To Use Java With Blogger: A Tutorial

Tue, 18 Apr 2006 16:36:00 +0000

Blogger has a REST API. I use it to retrieve particular posts or to post transformed data. There is no Java API that I know of, but you will see here it is not very difficult to interface with Blogger API in Java using plain old XML.

Using libraries commons-httpclient and DOM4J it would be quite easy to implement your own Java Blogger API as the following code will suggest.

Authenticate
All requests need to be authenticated and are done in HTTPS. I use common-httpclient to perform requests. Here is how to setup the client:

private  HttpClient initHttpClient()
{
  HttpClient client  = new HttpClient();
  List authPrefs =  new ArrayList(2);
  authPrefs.add(AuthPolicy.DIGEST );
  authPrefs.add(AuthPolicy.BASIC);
  client.getParams().setParameter (AuthPolicy.AUTH_SCHEME_PRIORITY, authPrefs);
  client.getParams().setAuthenticationPreemptive(true);
  Credentials defaultcreds = new UsernamePasswordCredentials(user, password); 
  client.getState().setCredentials(new AuthScope( "www.blogger.com", 443, AuthScope.ANY_REALM), defaultcreds);
  return client;          
}

Get Your Posts
To retrieve the posts, you just have to query the right url, and parse the XML response. I prefer to use DOM4J, because of its handy asXML() method to print a node as XML. For simplicity I use a Map to store an XML entry.

public Collection getPosts() throws  HttpException, IOException, ParserConfigurationException, SAXException, DocumentException 
{
  GetMethod get = new  GetMethod("https://www.blogger.com/atom" +"/"+blogId);
  int  statusCode = client.executeMethod(get);
  if   (statusCode !=  HttpStatus.SC_OK)
  {
    throw new RuntimeException(" Could not make HTTP  request properly: " +get.getStatusLine());
  }
  InputStream response  = get.getResponseBodyAsStream();
  SAXReader reader = new  SAXReader();
  Document doc = reader.read(response);
  Collection posts = new ArrayList();
  List entries  = doc.getRootElement().elements("entry");
  if (LOG.isDebugEnabled()) 
  {
    LOG.debug("found  "+entries.size()+"  entries");
  }
  for (int  i = 0; i <entries.size();i ++)
  {
    Element entry  = (Element) entries.get(i); 
    Map m = new  HashMap();
    for (Iterator it = entry.elementIterator();it. hasNext();)
    {
      Element detail =  (Element) it.next();
      String name = detail.getName();
      if (name.equals("link"))
      {
        m.put("link ",detail.attribute("href").getValue());
      }
      else if (name. equals("content"))
      {
        m.put("content",detail.asXML());
      }
      else
      { 
        m.put(name,detail.getTextTrim());
      } 
    }
    posts.add(m);                   
    if  (LOG.isDebugEnabled())
    {
      LOG.debug( "found="+m.get(" title")+", url= "+m.get("link")); 
    }
  }
  return  posts;
}

Create XML for a new Post
Nothing particular here, just XML production.

private String createXmlForCreatePost(String postTitle, String postContent) throws   IOException, DocumentException
{
  SAXReader xmlReader =  new SAXReader();
  xmlReader.setValidation(false );
  Document doc = DocumentHelper.createDocument();
  QName rootName  = DocumentHelper.createQName("entry", new  Namespace("", "http://purl.org/atom/ns# "));
  Element root = doc.addElement(rootName);
  Element title = root.addElement("title"); 
  title.addAttribute("mode"," escaped");
  title.addAttribute("type ","text/plain");
  title.setText (postTitle);
  Element generator = root.addElement("generator ");
  generator.addAttribute("url" , "http://31416.org");
  generator. setText("31416 Java Generator ");
  Element content = root.addElement("content ");
  content.addAttribute("type" , "application/xhtml+xml");
  //Element  div = content.addElement(DocumentHelper.createQName("div",new  Namespace("","http://www.w3.org/1999/xhtml")));
  //div.add(...); //YOUR XHTML HERE
  StringWriter result = new  StringWriter();
  XMLWriter writer = new XMLWriter (result);
  writer.write(doc);
  writer.close();
  return result.toString();
}

That's it

Last week Javablogs.com top 10

Tue, 18 Apr 2006 11:02:00 +0000

Web Services No Better Than CORBA?

Mon, 10 Apr 2006 15:23:00 +0000

Richard Monson-Haefel (RMH), author of the famous Enterprise Java Beans (O'Reilly) and of a Web Services book is making an alarming claim about JAX-WS, the new Java web services standard:

JAX-WS still sucks!

In an effort to write a client to real world web services (for Google, Amazon, Ebay), he mostly failed. Errors seem to be largely related to WSDL to Java conversion. Jason Green, a JBoss developer, managed to get quickly a web service working for Ebay, but analysis by RMH suggest that hundreds of classes were generated for this simple one method web service.

Recently I had to try to do CORBA client classes by hand in Java since the rmic tool was not working that well to generate classes that could be run on a client JVM of different version than the server JVM. It proved to be a no-go in the end, because of the complexity involved not only in coding, but more on maintaining those modified generated classes. This would have required coding a tool dedicated for that purpose. We chose to pass simpler objects, differently, which looks just like a hack to avoid having CORBA issues. I am amazed that after that many years of CORBA, there are still simple cases where it does not work properly automatically.

Seeing that there is the same kind of problem with Web Services (IDL to Java translation) makes me wonder about Web Services "improvement" over CORBA. And WSDL is definitely less readable than CORBA IDL.

Is your UML Composition French or American?

Fri, 07 Apr 2006 11:40:00 +0000

The French Way
In many French books/lessons (a lesson here, another from caen university , one from UQAM), UML composition does not seem to mean the same as their English counterparts. They write:

Si l'agrégat est détruit (ou copié), ses composants le sont aussi.

les composants (en tant qu'ils sont composants) ont une durée de vie incluse dans celle de leur composé.

translated to English:

If the aggregate is destroyed (or copied), its components are destroyed (or copied) as well.

components (while they are components) lifetime is included in the one of their aggregate

But then what does it mean to have a cardinality of 0 on the aggregate side?
Few French lessons give an answer (UQAM does, maybe because it's half French only). The component can be created before the aggregate but must be destroyed when the aggregate is.

The American Way
Most English books/lessons describe composition more accurately. Ootips definition is simple and clear:

Composition [...] is exactly like Aggregation except that the lifetime of the 'part' is controlled by the 'whole'. This control may be direct or transitive. That is, the 'whole' may take direct responsibility for creating or destroying the 'part', or it may accept an already created part, and later pass it on to some other whole that assumes responsibility for it.

However Martin Fowler in UML distilled explicitly writes:

The "no sharing" rule is the key to composition . Another assumption is that if you delete the polygon, it should automatically ensure that any owned Points also are deleted .

By using the word "assumption" he is avoiding a clear meaning. So I looked up in the UML specifications from uml.org. The specs are actually very near ootips definition and even more explicit:

An association may represent a composite aggregation (i.e., a whole/part relationship). Only binary associations can be aggregations. Composite aggregation is a strong form of aggregation that requires a part instance be included in at most one composite at a time. If a composite is deleted, all of its parts are normally deleted with it. Note that a part can (where allowed) be removed from a composite before the composite is deleted, and thus not be deleted as part of the composite. Compositions define transitive asymmetric relationships—their links form a directed, acyclic graph.

So it turns out that some French definitions were not wrong, but imprecise. Their true meaning becomes only clear after reading the English UML specs.

Ootips has also a good example of difference between aggregation and association. If you have an aggregation relation between two Node entities, it means you represent a tree of Nodes, not a graph of Nodes. See the link for more details.

Are UML Aggregation and Composition Good Concepts?

Tue, 04 Apr 2006 11:12:00 +0000

Yesterday a work collegue made the remark that the notions of aggregation and composition were not very clear. I was a bit surprised at first. But I actually almost always use just associations relations, sometimes composition, I was not sure anymore about the difference between aggregation and association and the one between composition and aggregation. I had to look up in a book to remember what it was all about.

The French book we looked at was really bad. In the end we did not understand fully the differences except a notion of "stronger coupling", some kind of "asymetry"!?! The difference between aggregation and composition was more explicit since in a composition, the "child" can only have one parent.

The UML Distilled book is much more explicit and I recommend that book to anybody doing UML:

Aggregation is strictly meaningless ; as a result, I recommend that you ignore it in your own diagrams .If you see it in other people's diagrams, you'll need to dig deeper to find out what they mean by it . Different authors and teams use it for very different purposes.
Composition is a good way of showing properties that own by value, properties to value objects (page 73), or properties that have a strong and somewhat exclusive ownership of particular other components.

Bertrand Meyer prefers to describe relationships using only two concepts, client and heir, basically UML association and specialization and therefore avoids the distinction between aggregation and composition and association:

B is a client of A if every object of type B may contain information about one or more objects of type A.
B is an heir of A if B denotes a specialized version of A.

He then uses "class invariants" to specify the nature of the client relationship.

SOA and B. Meyer

Fri, 31 Mar 2006 15:08:00 +0000

While rereading parts of Object Oriented Software Construction from B. Meyer, I still find valuable information I overlooked. In the part about functional decomposition where he argues pro and cons of top-down / bottom-up approaches, he has the intuition of the current SOA hype. By showing that

Real systems have no top.

He suggests that the only way to build complex software is through a service oriented architecture. Of course SOA is a very old concept. Meyer example of Operating Systems is a fine one.

Tired Of 'My 10 Best XXX' Blog Posts

Thu, 30 Mar 2006 10:32:00 +0000

It's incredible how many posts on popular aggregators are just about a list of stuff, and how except one word they sound the same.
Only today on Javaworld, I can read:
" Ten Things Every Java Developer Should Know About Unix"
"My 10 Favorite Scripting Languages "
"7 Must-Have Programs for Windows"

They all seem to have read the
"10 best ways to have your post popular on delicious"

It's a bit boring.

Bilinear Gaussian Lanczos? Downsampling!

Fri, 17 Mar 2006 17:32:00 +0000

Not easy to choose a downsampling algorithm. Some links:

mplayer: advises bilinear for quality and describes artifacts for each algo
a blogger: advises lanczos. This links shows pictures produced by each algo.
digital image interpolation: seems to have a preference for bicubic, what photoshop uses.
videohelp: tested various algorithms and think Bicubic spline is the best.

People don't seem to agree. If you want to check the theory:

Getting Started With DOM, XOM, DOM4J by Parsing an RSS Feed - An Experience Review

Fri, 17 Mar 2006 10:53:00 +0000

Recently, I looked for a way to get info from some particular blog entries of mine on blogger.com. Blogger used to offer an XML-RPC API . They even designed a version 2 of the XML-RPC API that does not seem to have ever been put in production. Or maybe I did not manage to make it work. I had no problem to make v1 work, however. I used apache XMLRPC v2, it was very simple to use. Unfortunately information accessible through XML-RPC Blogger API was incomplete for me. Furthermore, it is very likely that this API will disappear soon as it is deprecated since 2002.

Blogger wants you to use their Atom API. It is not XML RPC anymore, you have to do the parsing by hand.

The DOM Experience

I thought "no big deal, I will use DOM for it". I don't need performance and wanted a quick way to solve my problem, plus DOM does not require X number of libraries. It was easy to use regular DOM until I was frustrated by not being able to get the <content> element full text easily as it is sometimes XML. I did not want to hand code a method to do that as I thought it should be done by the XML library.

The XOM Experience

I heard previously of a simple XML parser, efficient, with an API that had been well designed, XOM. I looked at the API, there was a toXML () method to return the node content as XML (children included), sounded good. I saw there was even XPath support, and thought, great, it will simplify my code a bit. I will get the blog entries by just querying for " /feed/entry". No luck, it did not work, it returned 0 results. So I looked for mistakes in my code, did not find obvious ones. I tried other queries like " //feed/entry" or "//entry", same thing, not the right results. There must have been something wrong in my code, or maybe the XPath engine in XOM has particular settings to deal with RSS feeds (they contain various xmlns declarations). The point is that I got frustrated, it was supposed to be very simple, and in reality, not so!

The DOM4J Experience

I had experience with Dom4j before, just once, to build XML, not to parse it. I had relatively good memories of my Dom4j experience for that so I decided to try it out on my problem. At first I found Dom4j API a bit confusing as there are so many methods on most used classes. This is because Dom4j is DOM compatible. But I quickly understand the logic of it and found some very useful methods, namely Element.elements(name) to get all children elements by name. Of course, they have an asXML() method like XOM. There is also XPath support.
I tried the XPath on Blogger RSS without success again. There really must be a trick to get it to recognize RSS. But with the elements("entry") method, I very quickly got the same with not much more code, and it worked.

so DOM vs. XOM vs. DOM4J = 0 - 0 - 1

Example Code:


SAXReader reader = new SAXReader();
Document doc = reader.read(response);
Collection posts = new ArrayList();
List entries = doc.getRootElement().elements("entry");
if (LOG.isDebugEnabled()) 
{
        LOG.debug("found "+entries.size()+" entries");
}
for (int i = 0; i <entries.size();i++)
{
        Element entry = (Element) entries.get(i); 
        Map m = new HashMap();
        for (Iterator it = entry.elementIterator();it.hasNext();)
        {
                Element detail = (Element) it.next();
                String name = detail.getName();
                if (name.equals("link"))
                {
                        m.put("link",detail.attribute("href").getValue());
                }
                else if (name.equals("content"))
                {
                        m.put("content",detail.asXML());
                }
                else
                {
                        m.put(name,detail.getTextTrim());
                }
        }
        
        posts.add(m);                   
        if (LOG.isDebugEnabled())
        {
                LOG.debug("found="+m.get("title")+", url="+m.get("link"));
        }
}

Bad JDK Design Example

Thu, 16 Mar 2006 15:02:00 +0000

Have you ever used the Authenticator class? It is standard since JDK 1.2. This allows you to establish authenticated HTTP requests. OnJava has an in-depth article presenting it. Now here is my story about it.

The other day I just wanted to do that a simple authenticated HTTP request. You would expect it to be very simple to do in the modern internet world.

It is actually not very difficult to do, by hand, a JavaWorld article presents this way. Unfortunately it uses internal JDK sun classes for Digest and I am uncomfortable to use that since it is not guaranteed to work on any JVM. So I have to download jakarta commons-codec to just do a simple authenticated HTTP request. I really wonder why some of those codecs are not in the JDK (not as sun.* classes but as java.* classes).

Current Popular Subjects In Java Blogs

Wed, 15 Mar 2006 20:00:00 +0000

Based on my blogs entries and success of other blog entries, I present the current trendy subjects and the not trendy ones:

In:

Design Patterns
Ruby on Rails
EJB (bashing)
RSS
Spring
Hibernate
Javascript and AJAX
Java Virtual Machine

Out:

RMI
Portal
Jini

Neutral

JDK 5 features

As you can see most of the popular stuff is fairly common, not that new. I will do a finer, less common, analysis later if I have time.

Non Fiction Books (Mostly Software Programming) Before 2003

Wed, 15 Mar 2006 15:03:00 +0000

A page with useful reference for me as I don't own each one of these books. It can provides you a very short review non-commercial biased.

2003

Programming Jakarta Struts (O'Reilly - Chuck Cavaness) much better than Manning book. Explains struts very clearly with a more interesting example. Not enough on how to design your application for use with the evil ActionForm. Almost nothing on Struts 1.1 features. Nothing on common problems encountered while building web applications, or even about on how to design them using struts.
Struts In Action (Manning) not much more about struts than what's available on struts website. Their example throughout the book is the same login example. Struts 1.1 features, esp DynaForm are mentioned but not really presented. On the positive side, there is a very useful Validator reference at the end and a too short mention of different strategies for the ActionForm.
Contributing to Eclipse (Kent Beck - Erich Gamma) funny little drawings inside to make some good points.
The Design Patterns Java Companion(Adison Wesley - James W. Cooper) A must read for clean swing development. Essencial design patterns inside. Nice tip about doing deep cloning in java.
Java Threads (O'Reilly - Scott Oaks and Henry Wong) very clear presentation of multithreading programming in Java. Explain how to write your own thread scheduler.
J2EE expert one on one (Wrox) pragmatic view of J2EE. Many interesting comments like why avoiding stateful session bean if you are a little bit concerned with performance. Some strange benchmarks at the end of the book showing velocity templates being 10x faster than JSPs.
Java Distributed Computing(O'Reilly - Jim Farley) very good to learn more about RMI because it starts with a custom distributed object system.
Java Enterprise in a Nutshell (O' Reilly - David Flanagan, Jim Farley, William Crawford, and Kris Magnusson) good complementary book, provides a good basis. I liked how RMI use in EJBs was detailed, Servlet chapter is also excellent (but don't look for design tips/help in it - it is a book on the BASICS).
Enterprise Javabeans, second edition (O'Reilly - Richard Monson-Haefel) second lecture made me better appreciate that book, contains a lot of fundamental concepts to know. Very detailed on transaction management. However there is nothing I remember in it about EJB restrictions. In short incomplete but not bad.
Structure And Interpretation Of Computer Programs (MIT Press)

2002

How Mac Os X Made Me Discover a Bug in my Java Prog

Tue, 14 Mar 2006 20:15:00 +0000

Yesterday I had some free time to finally find out why MacOs X would not display my little applet properly. I checked various JDKs for MacOs, no difference. I checked if it was due to antialiasing use, no luck. I actually found out there was an error in the way I displayed images. I did not call repaint() between each image change. Strangely, it worked fine on Windows with many JDKs. Anyway now my Benham Disc Applet is working on Apple computers as well.

Del.icio.us Firefox Toolbar Update

Mon, 13 Mar 2006 19:45:00 +0000

I updated my little delicious toolbar that provides autosave feature for me. I did not bother reworking the new delicious extension as I don't use its new functionalities. However people are free to do so if they want with the code from this toolbar.

With Firefox 1.5, display of saved documents was not working properly anymore due to a change in Firefox handling of security permissions to view local files. It is fixed now.
Tags: delicious, toolbar, extension, firefox

Why did Netscape choose Javascript?!?

Thu, 09 Mar 2006 17:10:00 +0000

Last year, I have helped building an AJAX web application, where the web client was composed of only one web page and tons of javascript. The 1 web page for a site approach is similar to Gmail and many full AJAX sites. This approach looks quite elegant. Using JSON (and JSON templates) we were able to separated fully presentation logic and business logic and componentization seemed to occur naturally.

But I remember how often I complained on how bad Javascript was to code. I find it does not encourage good practices at all. The object orientation of Javascript is really a pain to use. If I had not a small experience with Python before, I would have thought scripting languages were really bad at building applications, even not so big ones. I find that even for good programmers, Javascript is a challenge, as it seems to always be a struggle to organize properly your code with it.

New Mac Mini

Tue, 07 Mar 2006 20:51:00 +0000

Here is my new Mac Mini Core Solo. I am quite happy with it since it is very quiet. It makes a good jukebox/movie/server machine. Apple was very quick in shipping it (2 days)

However I was a bit disappointed by iPhoto 6, it is not as nice as Picasa to sort out pictures (no IPTC support). I did not try Java on it yet. I just found out my little Benham circle applet was not working properly on it.

Features That Ruby Lacks and Java Has

Fri, 10 Feb 2006 15:57:00 +0000

Bertrand Meyer describes in his book " Object Oriented Software Construction (2nd Ed)" qualities that a good object oriented language should have. Java has almost all the features. Ruby is much farther off, which does not mean it is not a good language, I think it's very good to write some types of programs quickly (should I call them scripts?), syntax is quite nice, but it does lack the following useful features:

Assertions:
The language should make it possible to equip a class and its features with assertions (preconditions, postconditions and invariants), relying on tools to produce documentation out of these assertions and, optionally, monitor them at run time.
Information Hiding (Java is not good either on that one, the protected keyword is of dubious value):
It should be possible for the author of a class to specify that a feature is available to all clients, to no client, or to specified clients
Static Typing:
A well-defined type system should, by enforcing a number of type declaration and compatibility rules, guarantee the run-time type safety of the systems it accepts.

Genericity:
It should be possible to write classes with formal generic parameters representing arbitrary types.

More pragmatically, library choice and performance difference between Java and Ruby might be the most decisive.

Tags: book review

What Is 'Modern' Java Compilation?

Tue, 07 Feb 2006 11:12:00 +0000

Occasionally in Ant you can see messages like this:
"[javac] Using modern compiler"

What does this mean?

In Ant you have the property " build.compiler" to specify if you want to use a classic or modern compiler. Now what do they mean by modern or classic. Well, they call classic compiler the compilers of JDK 1.1 and 1.2 and they call modern compiler the ones of JDK 1.3+. They made that distinction because a classic compiler does not support the same options as modern compilers: the semantics of javac tool changed in JDK 1.3.

This terminology can easily be confused with java class file compatibility. Java class file compatibility is changed using the "-target" option of javac tool. One can specify in ant to compile with modern compilers and a target 1.4, the result is likely to not run on JVM 1.3. One can specify modern and a target 1.1, the resulting classes will run on JVM 1.1, but ant build.xml file will not be usable with JDK 1.1 without changing modern to classic.

I find Ant choice of word a bit confusing as using a "modern" compiler has little to do with the resulting class files.

RMI Nightmares - Non Interoperability

Tue, 31 Jan 2006 18:39:00 +0000

These days we tried to make Websphere and Jonas interoperate. Some EJBs are deployed in Jonas and we just try to call them from Websphere. The most natural way is to use RMI (with JRMP) for that. Jonas uses JDK1.4.2 (conforms to J2EE 1.4) and Websphere IBM JVM 1.3.1 (conforms to J2EE 1.3). Because of RMI serialization, an EJB from Jonas will not work on Websphere, because this EJB has references to J2EE 1.4 classes which are different from J2EE 1.3. The exact root cause is not exactly clear to me, it probably has to do with the Naming Factory and the EJB Stub, but we tried every possible thing here without success.
Update: JRMP worked between JBoss 3.X on JDK 1.4 and Websphere 5.X (IBM JDK 1.3) using JBoss NamingFactory. So RMI interoperation seems quite unpredictable. One thing is that you have to pay attention to the naming factory you use. If it is a special library like in JBoss case, make sure it is compatible with your JDK.

One solution is to use RMI-IIOP, as it does not rely on serialization. It almost works. In our case we use Websphere Portal that needs to have security enabled. You might ask what security has to do with all that. Well, enabling security on websphere means that CORBA exchange will also be secured (using CSIv2). Tuning Jonas to understand what websphere asks is something we did not manage to do (and we don't know if it is possible at all).
Update: We did manage to make it work by disabling Jonas CSIv2 support.

So there we are using Axis to do the communication!
Update: Axis is not needed anymore as RMI-IIOP works as advertised.

There is an interesting benchmark of Axis vs JRMP vs RMI-IIOP (pdf). Axis is more than 10x slower than pure JRMP. Application servers use of JRMP seems to matter a lot, JBoss looks bad.

What Is Clustering About – An answer to Spring is not designed for scalability

Mon, 16 Jan 2006 16:57:00 +0000

I wanted to reply to Diego Parilla post about scalability, but my reply ended up being too long not to post it over here. I don’t know if I am the only one here but I don’t fully understand his post and arguments.

Clustering the web server and clustering the model are just two parts of clustering, not exclusive.

In a LAMP application, clustering the model is traditionally done by clustering the DB (MySQL) which is a no brainer. Clustering the web server does not always implies sharing the “HTTP session”. For example you can limit yourself to use cookies. For many web apps this works well.

My Dream Job

Thu, 12 Jan 2006 11:42:00 +0000

In my dream job, I would have freedom to play with and build software based on new ideas I have.
In my dream job, I would work with creative people, luminaries.
In my dream job, I would have a good enough salary.
In my dream job, I would not have to worry about the company stability.
In my dream job, I would have freedom in my work schedule.

Get paid for doing what you really want (i.e. your ideas) is possible. But without worrying about the next 6 months, I have not seen that.

Ogg better than MPC and MP3 - AAC crap

Wed, 04 Jan 2006 11:47:00 +0000

The latest well done (e.g. blind) listening test comparing Ogg, MPC, MP3, AAC at about 180kbps shows that Ogg is superior to all other compressing formats. It’s not that easy to find relevant tests. Many people claim to be able to notice huge differences between CD and MP3 and yet, but confronted to a blind listening test, they are very surprised how difficult it actually is. Hydrogenaudio has a good community of audiophiles.

What's Your Favorite Design Pattern

Tue, 27 Dec 2005 12:18:00 +0000

Once in a job interview, someone asked me

What’s your favorite design pattern?

I was first surprised at the question. Usually a design pattern is about solving a problem not about a personal preference. I wondered if it was a trick question of some sorts. I was hesitating with my answer and asked more about the meaning of the question. I just had to really give the pattern I liked the most! I really should not have but I ended up pleasing my interviewer and said “the decorator pattern”. At least its name is somehow related to arts.

Professional Portal Development Book Review

Tue, 27 Dec 2005 12:08:00 +0000

I just finished reading Professional Portal Development With Open Source Tools (Wiley).

The Good:

I enjoyed Chapter 1, it gives a good presentation of Portlets for someone who does not know much about it. It is a chapter one can use as a reference to verify some specific details of how Portlets are working. It is a bit similar to JSR 168 specifications, but I found it more detailed and well written.

IBM Websphere Portal Strategy

Wed, 07 Dec 2005 17:48:00 +0000

I digged a bit deeper in a current project where they use IBM Websphere Portal Server (WPS) with IBM Web Content Management (WCM). I found out that both products collide and disintegrate.

WCM is like a regular content management system, quite flexible and performant on its own. It has a nice way to separate layout concern from content concern. It has a flexible cache. It has a user repository to set different permissions. Its architecture is even so well done (very modular) that you can tailor it to your needs to do things a content management system is not supposed to do. It could do most of what you would do with a Portal system. And that’s the problem.

Portlets in an AJAX World

Tue, 29 Nov 2005 16:29:00 +0000

Recently, I have had my first encounter with Portlets although Portlets are not very new. IBM had their own Portlet API before the JSR 168 was final in 2003. Unfortunately both approaches are quite different and IBM does not recommend the use of their own API.

Portlets were a needed concept for a long time. I worked in 2000 on a java portal system and it would have been handy at that time. Portals are everywhere since the internet bubble.

But Portlets do not do that much either. Sites designed with strict and intelligent rules can bypass the portlets need for example by using a combination of custom tags and java beans.

What shocks me a bit more is that while not doing much, it looks a bit old fashioned already. With AJAX you could build very flexible and user-friendly Portlets replacement. This AJAX-let is actually probably what Microsoft has already done on their start.com web site. Now with AJAX, I just wish Javascript would evolve (a lot) or die to be more OO friendly.

Good Software Books

Thu, 17 Nov 2005 21:24:00 +0000

A friend asked me for a list, here is it. I found those books the most interesting ones among those I have read about software in the past year or two:

Artificial Intelligence through Prolog, by Neil C. Rowe (Prentice-Hall): if you don't remember much about Prolog, it's a good book. It details how Prolog interpreters work.
File Systems Forensic Analysis, by Brian Carrier (Addison Wesley): everything you need to know about disks and file systems, every trick to recover lost data. To go more in depth into file systems theory, you can read the free Practical File System Design with the Be File System from D. Giampaolo, the creator of BeFS.
Mind Hacks (O'Reilly): you probably already bought that one. Not about software, but excellent.
Practical Issues in Database Management by Fabian Pascal (Addison Wesley): short but concise book on relational database theory. F Pascal is of CJ Date school of thoughts, "null are evil". You don't need to read it if you have read any other similar book (by CJ Date for example).
Inside The Java Virtual Machine, by Bill Venners (McGraw-Hill): there are other books on the same subject. While this book has some bad reviews, I found it an easy read, and it explain well enough for me all the inner details of Java. I found "Programming for the Java Virtual Machine" not better, and with some bad example of a Prolog language for the JVM (1 full chapter for this to present a way too simple thing to be of any use).
Programming Jabber, by O'Reilly: I have read it 2 years ago, I enjoyed how they made you go through building a Jabber server.
Lucene In Action: the only book about Lucene, some of the info is useful to understand how it is designed.
Hibernate In Action: if you have to use hibernate, this is the book to read. It presents different ways of using it.
some Javascript+DHTML(+CSS) book: useful if you have none. They often have good tricks to solve the usual problems. In the AJAX days, it becomes useful again. I find O'Reilly's "Javascript - The Definitive Guide" good to have an API reference (as ebook), and "Javascript and DHTML Cookbook" has the most useful recipes.
Lean Software Development by Poppendieck (Addison Wesley): easy to read, good presentation of software dev management problems and pragmatic solutions. My favorite on the subject. Another good one but very specific is "Requirements Management" (MS Press).
UML Distilled by M Fowler: always useful to have if you need to draw UML diagrams.

Good luck!

Apple Rumors and X86 first move

Wed, 16 Nov 2005 17:59:00 +0000

I find the mac mini a particularly compelling computer. It is small, noise friendly, relatively cheap. If you want to have a cheap home server (based on Linux or MacOS X), this is quite a good choice as you probably don’t want to hear all the fans of bigger computers the whole day. It’s also a very good candidate for a media box: you can connect a huge hard drive for mp3s, and watch your DVDs, photos on a nice LCD (and MacOS helps here).

7 Years Later: Still EJBs

Mon, 07 Nov 2005 16:29:00 +0000

In 1998, a student of my school had a project with Sun about EJBs. This was the beginning of EJBs, they were not public, and there was no J2EE. At that time he found it utterly complicated, and had difficulties to understand why they made it so complicated.

In 2000, I worked for a company who had built in-house most components of a today J2EE application server. For example, they had their own unified messaging system, on the JSP side, they were already using custom tags, etc.. Their problem is that they took lots of time to develop in-house, all what later became standard, and almost freely available.

Today, 7 years later, the big talk, new Java products, be it Spring, or JBoss Seam, are still all about the same ideas as in the first J2EE. Spring has the freedom to include whatever technology, but in the end, their scope closely matches J2EE scope. Service Oriented Architecture (SOA), the current buzzword, will come out relatively naturally with J2EE, session EJBs being natural candidate to expose functionalities to the outer world. JBoss (or Geronimo) architecture is fully SOA: a small kernel managing services, that are exposed for other to use.

To learn about EJBs, I still find Enterprise JavaBeans by Richard Monson-Haefel, the best book. The chapter (in the 2nd Edition of the book, not in the 4th unfortunately) where he rolls his own distributed objects system is for me a testimony on how well he wants you to understand what's behind EJBs. And if you want to better understand what's behind RMI, Java Distributed Computing (O'Reilly) has a chapter where the author rolls his own RMI.

To conclude, Sun was really clever in defining J2EE scope and branding several Java technologies as J2EE (despite the awful version naming). While the ideas were very good, the developer's life was forgotten. It was too much a pain in the a** and not flexible enough (especially the ORM side). JBoss gets it right, what people need is a way to use pieces of J2EE, without imposing too much unnecessary hurdle on development.

Job Hunting: Interview Madness

Fri, 04 Nov 2005 10:40:00 +0000

I have been doing interviews those past weeks. Today I had to present my resume for the N time where N is big. It is very boring. In France, it is common to have 5 interviews for a small company (that says a bit about how people trust each other and makes me envy Germany where the practice is to select very strictly on the resume and be more relax about interviews). But all those interviews give a valuable insight on the current market, and on some companies.

Does Eclipse Rot The Brain?

Thu, 27 Oct 2005 17:42:00 +0000

Slashdot presented a really insightful article on Visual Studio by Charles Petzold called Does Visual Studio Rot the Mind. Interestingly it seems that Eclipse does things better than the future Visual Studio.

Autocomplete in Eclipse is really good and does not suffer of the bottom-up problem M. Petzold mentioned, nor from the forced CTRL-Z.
There is no Form designer, and in Java it has been the practice to lay out forms programmatically. The Java Layouts have always been scalable, and Java programs very rarely rely on pixels.
There is no unnecessary import statements.

His positive feedback on XAML is interesting, especially since he thinks this allows for a better automatic UI.

Orkut Statistics and Blog Evolution

Tue, 18 Oct 2005 13:03:00 +0000

I was just having a look at orkut again. As in my last look in the past year, there is not much new or interesting. The technical forums are mostly uninteresting and it almost looks like nothing is going on. Take a look at the orkut java forums, in 1 year, very few messages are good.

Compare that to javablogs.com blogs aggregation, every week there are many interesting posts. The blog model is good because it is targeted at people who want to write. When you blog, you build a history. Forums don’t let you do that. You could have a social model on top of blogs, and I am sure this will be one of their next evolution.

Spring Books Roundup Updated

Mon, 10 Oct 2005 16:27:00 +0000

I am completing my Spring Books Roundup with 2 new books:

Pro Spring, by Apress: a lengthy book, with some unnecessary information, but good overall. I did not like the Part 1, there is not much content in it. Fortunately, Part 2 is much better, dependency injection is very well described, almost on par, with the Fowler article of reference. Other Spring areas are well covered, in a similar manner as in Wrox book, except Java Server Faces, absent from the book, and nothing on Swing as well
Spring Reference Documentation, from the springframework.org website: I should have started with this one, it is of very good quality, and more up-to-date. There is notably a very good chapter on “Source Level Metadata Support” even if I don’t find particularly compelling that they chose yet another abstraction behind Java 5 annotations, partly to support older JDKs. Another plus is a chapter on JMX support, this was lacking in other books.

Interestingly most other books are very similar to the reference documentation, in their presentation and content. The reference documentation is extremely well written. Each time the author carefully explains what are the advantages of the Spring solution, or what does a particular feature solve. Only dependency injection and swing are not treated properly.

Spring Books Roundup

Mon, 10 Oct 2005 11:15:00 +0000

Spring is hip these days, so I decided to learn a bit more about it. I had used Avalon a while ago, I was attracted by its design by component and the way it seemed to lay out a proper infrastructure to build a server application. In the end, I was a bit disappointed, it required a bit too much glue code for my taste and did not provide that much in exchange. I don’t remember if, at that time, it was advertised as IoC (Inversion Of Control) container. This experience is one of the reasons why I did not jump on the Spring ship.

Interesting Plug-In Framework - DPML Transit

Thu, 29 Sep 2005 15:20:00 +0000

Today, I just found out about DPML Transit, it is a small framework that helps you build plug-ins based software. It seems to work a bit with DPML Magic, their build system based upon Ant. Both are quite interesting, since in big projects, you often end up with a packaging per component (which DPML Magic seems to make very simple) and a versioning of those components. DPML Transit allows then for an efficient way to look up a particular version of one component.

Java Puzzlers - Can you figure this out?

Wed, 28 Sep 2005 16:04:00 +0000

The book Java Puzzlers is quite good. I don't think anyone can get every puzzle right. This shows again how you can very easily make someone fail interviews if you ask too silly questions. I suppose that if people were asking those questions they would not expect the right answers, but study the candidate reactions.

Here is a sample:

public class DosEquis { public static void main(String[] args) { char x = 'X'; int i = 0; System.out.print(true ? x : 0); System.out.print(false ? i : x); } }

This will output "X88". Obviously this is not good code, which is precisely one of the book objectives: to show how bad some practices can be. But at the same time you learn a bit more about the Java language and its possibilities. In the latter chapters they have more interesting puzzles.

Is Prolog Better Suited Than SQL?

Mon, 26 Sep 2005 15:36:00 +0000

I am currently reading a Prolog book Artificial Intelligence Through Prolog, I have been doing a bit of Prolog when I was very young and wanted to refresh my memory a bit. It is a very interesting read, especially when I take the viewpoint of our current application where no ACID compliance is required.

It seems to me that all the logic we coded to parametrize SQL queries and construct them dynamically could have been avoided if we had chosen Prolog as Prolog expressions would have been very natural to use in our project. With Prolog, there is no need to think about joins, type of joins, SQL syntax. It is at the level just higher. I wonder very much why Prolog did not become more mainstream as it seems to solve some problems in a much nicer, natural way.

When California Was An Island

Wed, 21 Sep 2005 19:01:00 +0000

This is from an old (1680s) big (2 tons) Coronelli globe, currently displayed in Grand Palais, Paris. Can you spot San Francisco there?

Currying - A Very Interesting Use Of Generics

Fri, 16 Sep 2005 14:41:00 +0000

In bamboo-dht, a distributed java hashtable project, the main developer, Sean C. Rhea, advocates the use of Curries and Thunks (of LISP and ML). He wrote an interesting and valuable document, Async Tutorial, presenting a use of it. I will present the concept here shortly:

public interface Thunk1<T> { void run (T t ) ; } Thunk1<Integer> intThunk = new Thunk1<Integer>() { void run (Integer i ) { System.out.println( i ) ; } } public static <T> Runnable curry(final Thunk1<T> f ,final T t ) { return new Runnable () { public void run ( ) { f.run ( t ) ; } } ; } Runnable print42 = curry( intThunk , new Integer ( 42 ) ) ; print42.run ( ) ; //prints 42 to standard output

So it is a very clever way to have a callback in java. Very clever because you can declare your code to take only a run(), and pass any parameter in it by using a curry. It is very simple to use, even if the small framework around it can scare some people.

However neat the idea is, I am not sure it is practically useful. Sean C. Rhea used that because he did not want to use "unnecessary" instance variables. Ok, but the curry is still creating another instance (actually 2 but one could be static), is the overhead of a class instance without variables that much more than one with variables? And there is another way, without the Curry framework:
Runnable print42 = new Runnable({ public void run() {intThunk.run(new Integer(42));}});
print42.run();

This is actually exactly what the Curry framework would do behind the scenes. The Curry framework makes it a bit more elegant, but I am not sure if that's really more readable for most programmers. I would personally advise the traditional way, use instance variables.

On J2EE Portability Accross Application Servers

Tue, 13 Sep 2005 16:50:00 +0000

Brian Alexander Lee wrote:

“dismal interview where I asked about the candidate’s experience with porting a J2EE application from WebLogic to WebSphere (which was listed on his resume). The candidate said that it was very easy and he just deployed his application with no problems or changes necessary, he had no changes made for the application to run properly. This was one of many bad signs for the candidate.”

I disagree. I would even use his example to show that the portability game is more about configuration and packaging rather than about anything else so that many developers would in-deed not see a big deal into porting an application to a new application server.

JavaBlogs Weekly Top 10 and Java HTML parsing

Mon, 12 Sep 2005 16:00:00 +0000

I took some time to continue my little JavaBlogs analysis, I now have a page summarizing the top 10 most read blog entries in the last week. The page is generated every 24h (this is why there is no ‘best progression’ as of today).

I also fixed some bugs related to HTML in RSS2. I understand a bit better why a RSS 1.0 co-author decided to remove the possibility of HTML descriptions for RSS 3.0. It often does not make sense to keep all that information about styles, fonts, etc. from different sources. What I do is rewrite the HTML, allowing only b,i,a,p,br tags, with the style information stripped. I found the open-source htmlparser java library quite helpful to achieve that.

Spam In Blog Comments

Sat, 10 Sep 2005 01:34:00 +0000

I was a victim like many other of spams in comments. It’s stupid for people to do that on Blogger.com since the links on comments can not be referenced by search engines (they have some special ‘relative’attribute for that) and improve pagerank.

Fortunately Blogger.com provides a word verification step if you want to avoid random spam. However I am a bit disappointed that they force Blogger.com users to do that word verification as well. This time I find it stupid from Blogger.com. They have control on their users, so they could ban spamming users, and for everybody else on Blogger.com, this would be just one less step. I am always a bit annoyed at measures that solve a problem caused by a hand of people by making it more annoying for the majority.

JavaBlogs Daily Analysis

Fri, 09 Sep 2005 17:10:00 +0000

I was wondering what blog entries were the most interesting on Javablogs. I decided to write a small application to do that. It was not much more complex to put it online for others to look at as well. It is currently running on http://gopix.net:8081/javabuzz

It also presents Javablogs a bit differently (I like it better that way).

Please note that it is just the result of a 1 (full) day of work currently. I hopefully will have a bit of time to improve it. For example I’d like to add some graphs about popularity, some weekly stats, and comments in blog entries.

Commons-Beanutils Is Slow

Wed, 07 Sep 2005 17:28:00 +0000

The BeanUtil.popupate(bean,map) can be very handy, but if you care about performance, it is quite slow. I ran a micro benchmark on my machine (centrino 1.8ghz, JDK1.5) and found out that BeanUtils is up to 40x slower than a hand coded solution (where I assign each bean field manually). I was a bit surprised to find such a difference. I suppose there is a big penalty for using reflection and another big one for the BeanUtils abstraction (automatic casting, etc.). I did another test without BeanUtils, using if/else statements vs HashMap.get and found out that the if/else string.equals(…) statements can degrade performance by about 10x. The HashMap appears to be very performant, even with just a few elements in.

Generate your RSS feed in Java

Mon, 05 Sep 2005 14:03:00 +0000

There are some open source projects that can help you in generating or reading RSS feeds in Java. I found only two libraries a bit mature, other code is often embedded in other open source products (jroller for example):

Informa: Does various RSS formats and Atom 0.3. Documentation is better than its alternative, but less focused (has some hibernate helper thingy, some lucene helper, etc.).
Sandler: There is no working homepage while I am writing this. But the code is of decent quality, supports Atom 0.3 and RSS 1.0. It is easy to use it. However in reality it is not much more than a wrapper around some XML parser specialized in generating an RSS structure or an Atom structure.
Ooops, I forgot another important one, Rome. This RSS/Atom framework with a catchy name is very similar to Informa, has good documentation and good looking code. Under the hood it makes use of jdom.

I personally use dom4j since I only need to generate RSS, and RSS, or Atom are just XML. I don't find it particularly verbose to use dom4j for that, and it is very flexible.

If you need to parse feeds, then those libraries might make sense and save you a bit of time. For generating, I think their main interest is to abstract you from the differences in formats. So if you need to handle different formats, a framework will allow you to do it through only one API, which can be a big time-saver.

The Evil Port 80

Thu, 01 Sep 2005 13:46:00 +0000

I was writing an Atom feed generator for my current project. I chosed to support Atom 1.0 since it looks like it has the capabilities to establish as the next standard. Unfortunately I quickly saw that it was quite hard to test it in the real world (out of the good feedvalidator), as almost nobody seems to accept Atom 1.0 feeds yet, even if it is rapidely changing (there is support for it in Firefox CVS version).

So I decided to support RSS as well, the big question was: which RSS version? After grabbing lots of info on the subject, I opted for 1.0 again (more flexible, more different than Atom). It was actually quick to support RSS, but then when in real world, neither Google Desktop nor My Yahoo was willing to accept my feed. I looked at every bit of my xml, fiddled with Tomcat configuration in any possible way when I saw that no request was coming to my server from Yahoo or Google. And finally I thought, hmm maybe it's the port. I restarted my server on port 80, and yup, it worked!

I wonder why Google Desktop and My Yahoo don't support another port than port 80 for RSS feeds.

I Need Another DB Framework!

Tue, 23 Aug 2005 19:37:00 +0000

I am currently facing a problem that neither Hibernate nor iBatis solves nicely. I also looked at other ORM or just DB framework, without success.

What I would need is a framework that generates PreparedStatements with a query by Criteria like API. I have many queries that are similar but varying according to different input parameters. iBatis can handle this, but for complex queries and scenarios, the XML becomes completely unreadable, and you therefore loose any advantage that iBatis was bringing with the externalization of SQL statements in XML. The other issue I have with using iBatis is that for another part of my project, the automatic generation of SQL statements a-la Hibernate is useful. Hibernate has a very nice Query by Criteria API, but it lacks just a tiny bit of flexibility in customizing queries. For example, I could not find a way to specify a “USE INDEX(index_name)” in the generated SQL, after the SELECT FROM xxx and before the rest of the query. I did not find either a way to specify the use of a “STRAIGHT_JOIN” instead of an INNER JOIN. These are all MySQL specific issues, but those little things are extremely useful at improving some of my queries performances. Writing N sql queries hard coded is not a good option, since this N can be quite big, which is why I am using Query by Criteria in the first place.

del.icio.us toolbar customized

Sun, 21 Aug 2005 05:20:00 +0000

When I bookmark articles with delicious, I like to keep the content on my hard drive, because pages sometimes change, or are removed, or I want to do local searches. I believe this is one reason some people like furl (furl keeps a copy on their server that only yourself can read, but does not allow search).

A combination of slogger and delicious could solve partially the problem. But it is not integrated, I can't get my local version from delicious, so I loose the tagging, listing and all other plus from delicious.

I added my own feature to the delicious toolbar, which I like very much. This new toolbar saves automatically the file you bookmark (on the + button), and will add a link in your delicious home to the local version (if it exists).

I have it publicly accessible at http://perso.wanadoo.fr/logos0 1/deltoolbar.html

It is not meant to be used by everybody as it is not official. But the page will give you an idea of what it does. If you think it is useful, I will improve it, otherwise it will stay the way it is because it fits my use.

Categories: delicious, firefox, toolbar

Inside the Java Virtual Machine

Thu, 18 Aug 2005 11:03:00 +0000

I am reading an old book, Inside the Java Virtual Machine. Some old books don't age, and this is one of them. The chapter on the Java Virtual Machine is just excellent and should be read by every Java developer. It explains each step a JVM does when you run a Java program, very clearly.

You could get plenty of stupid interview questions from it like: How is the Java stack used? Between method area, heap, pc register, stack which one are shared among threads?

Also they saw the full potential of Java quite early on (1997). They explain how the JVM specs allow for very different implementations, ones that can run in different environments, for example, simplifying a bit: low memory, embedded world, or lots of memory, mainframe world. It is not an accident if Microsoft chosed a very similar design for the CLI of .NET, they have been looking for getting into the embedded area for quite some time, and apparently, they are making good progress.

Firefox Shortcut URL as Command Line for the Web

Tue, 16 Aug 2005 13:42:00 +0000

I have been doing some work on Firefox extensions. Mozilla/Firefox is a really interesting world, it is extremely flexible, you can make it transform the web the way you want.

Recently, someone started YubNub, the "command line for the web", a very simple idea, but useful. This is a bit similar to the Firefox Search Bar, except more powerful, and quicker to use. Mozilla developers actually had that idea a long time ago (around 2000), it is called shortcut url. Basically to any bookmark, you can add a shortcut, you then just have to type that shortcut (let say g for google.com) in the address bar to go to the address pointed by the shortcut. This shortcut accepts 1 parameter (represented by %s in the URL), so you can do "g chasethedevil" to search google for "chasethedevil" if you set up your shortcut properly. Firefox can build this dynamic URL automatically for you if in an input box, you right click and choose "Add a keyword for this search".

Now we have seen the Firefox address bar can be used to quickly perform a search on a selected website. But this can be much more powerful. As this page suggests, you can associate a bookmarklet with a shortcut, you can execute whatever javascript you wish. One very simple example is:
javascript:void(location.href= 'http://web.archive.org/web/*dc_/'+location.href). This shows how you can use the address bar in a very flexible manner. It would be easy to support a multiple parameters search using a bookmarklet. There is a whole list of commands very similar to yubnub ones in addition to those on the previously mentioned link.

I first thought about writing a simple extension to have my command line in Firefox (for example by extending the getShortcutOnURI function), but after some information gathering, I discovered it was already possible. Firefox is truly amazing. It is a bit of a shame that this functionality is not a bit better exposed to the user.

tags: Firefox shortcut extension bookmarklet

Hibernate Criteria API vs HQL Overhead

Wed, 10 Aug 2005 14:16:00 +0000

I was wondering what was the framework weight in the performance of my application, and I wanted to check at several frameworks performance. I used JMeter to benchmark a stripped application (database access through tomcat) under various loads. My very modest test shows no significant difference between Hibernate Criteria API and HQL access for queries. Criteria might be a few milliseconds slower, but my query time will take 10x more time, even for a relatively simple query (my query is has 3 inner joins, is grouped, with a count and takes only 4ms when performed once in mysql under no load). It seems much more important for the Framework of choice to provide good scalability, easy development and maintenance, rather than saving a few cycles, unless the Framework becomes a bottleneck.

How to become a JavaRSS member on blogger.com

Fri, 05 Aug 2005 13:32:00 +0000

The trick is to create a post with the javarss image and to use the link generated by blogger.com to display the image. Then you can follow the rules from http://www.javarss.com/pages/?page=suggest

Java is more productive than Ruby/Rails

Tue, 26 Jul 2005 13:26:00 +0000

I have been doing some Ruby On Rails, for 2 small projects. While I think it is good, I think it is overhyped as well. It is well designed, has good ideas (easy configuration), and focus on the right problem, architecture. But my conclusion is that I am not more productive with it than with Java.

I think most of the development time is not spent coding, but thinking. It is a very obvious statement, and yet too often ignored.

féminisme et statistiques

Thu, 21 Jul 2005 12:16:00 +0000

Sur ce lien, un article un peu trop long, mais avec pas mal de points intéressants:

6 femmes par mois décèdent, en France, par violence conjugale, soit plus que le nombre de victimes du metro anglais sur un an. Pour toutes les statistiques sur les chances de mourir suivant les causes, aux Etats-Unis, cliquer ici (on voit notamment que le plus grand risque est l'accident de voiture, ou la chute, ou le suicide par arme à feu, devant les attaques par arme à feu).
Un petit rappel sur l'antiféminisme chrétien, les femmes y étant qualifiées: de «défectuosité naturelle», de «raté de la création», d’«animal inquiet», de «mouche éphémère», de «méchante bête», de «porte du Diable», de «sac de fiente», de «diable domestique», de «racine du Mal», de «rejeton de tous les vices», de «piège tendu par l’ennemi», de «terre puante», d’«appendice de la race humaine». Ce qui lui permet d'ajouter: «Au lieu d’expliquer la misogynie des hommes par celle des religions, ne faudrait-il pas plutôt expliquer la misogynie des religions par celle des hommes?».
Une citation intéressante sur le voile en France: «ces filles qui affichent leur voile dans une société laïque sont le pendant exact de celles qui le refusent dans une société musulmane» laisse à réfléchir; et aussi un développement sur Theo Van Gogh dépeint ici en tant que raciste, misogyne et opportuniste, un avis que je ne partage pas.

C'est un peu extrême par moments, mais souvent juste. Sur le même sujet, Le Monde avait un petit article sur les wagons réservés aux femmes à Tokyo. Bon, j'ai peut-être un peu pété les plombs avec toute cette lecture.

Tags: feminism francais actualités.

Ruby, Python, JavaScript, Lua, Java, C++ benchmark

Wed, 20 Jul 2005 11:49:00 +0000

Simple benchmark but with interesting comments here:

Some of the conclusions are:

Java interpreter is very good
Java JIT is as fast as C++
Ruby, JavaScript, Python are of similar order of magnitude speed, Ruby the slowest.

It would have been interesting to benchmark interpreted languages for the JVM.

Now for most projects, architecture, not language is key in achieving good performance.

Fowler On Your Future: editing the abstract representation of a program

Tue, 19 Jul 2005 12:40:00 +0000

This Fowler article on "language workbench" reminds me of Gosling Jackpot ideas: it is about a different interaction with your program where the source code does not matter so much, but its structure. It must be fascinating to work on this kind of project.

Maybe MDA sounds similar to you, after all, actual language workbench systems are generating source code. But the fundamental difference is that a workbench generated program is much richer and can capture everything you want to do (MDA is too based on UML to be that powerful), also, the source could disappear and be replaced by byte codes only.

Now this all seems very nice, but unfortunately, there are always problems, and Fowler does not miss them: vendor lock-in (not that big of a deal if source is generated), integration with dev tools (especially version control), maturity. It looks like an open source language workbench would help here.

It is interesting to see that Intellij (yes, the company behind IDEA) is already well on its way. You can check it out in Fowler examples. His examples are a bit too simple and not very far off rules in rules engines, but the language definition tool does look very neat.

As Rainer Joswig shows in his video, LISP seems like a particularly good language to do this. This is hardly surprising since LISP is very near the abstract tree structure and therefore already very used in the AI circles.

Tags: java programming

Jython instead of Ant: sample code

Sat, 16 Jul 2005 01:50:00 +0000

Here is an example of how to use jython to deploy a web application under tomcat. The application is supposed here to be WAR packaged in a 'web' directory. Libraries used for the web-app are already in web/WEB-INF/lib, classes are compiled automatically by Eclipse under web/WEB-INF/classes, and resources are copied automatically by Eclipse. If necessary, it is very easy to script those actions in the python script, with very little code.

The longest code is the TreeCopier class, that should probably be part of any python make script. The reload method could be reworked to provide a generic method for reloading a tomcat web application. The following is just a working sample code as proof of concept.


# call jython make.py method1 method2, ... 
# to invoke method1 and then method2, ...

import sys, os, os.path, shutil, urllib, base64, re
web_dir = 'web'
webapp_name = 'myrestaurant'
tomcat_dir = 'c:/java/Tomcat 5.5'

#copy web directory to tomcat app dir
def deploy():
   target_dir = tomcat_dir+'/webapps'+'/'+webapp_name
   copier=LatestTreeCopier()
   copier.copytree(web_dir,target_dir)
   print "copied %s file(s) to %s" % (copier.count,target_dir)

#reload tomcat web application
def reload():
   url = "http://localhost:8080/manager/reload?path=/"+webapp_name
   base64string = base64.encodestring("%s:%s"%('michelin','michelin'))[:-1]
   opener = urllib.URLopener()
   opener.addheader('Authorization',"Basic %s"%base64string)
   h = opener.open(url)
   print h.read()
   h.close()

class TreeCopier:
   def __init__(self):
       self.count = 0

   def filter(self, srcname, dstname):
       return true

   def filterDir(self, srcname):
       return true

   def copytree(self, src, dst, symlinks=0):
       """Recursively copy a directory tree using copy2().
        """
       names = os.listdir(src)
       if not os.path.exists(dst):
           os.mkdir(dst)
       count = 0
       for name in names:
           srcname = os.path.join(src, name)
           dstname = os.path.join(dst, name)
           try:
               if symlinks and os.path.islink(srcname):
                   linkto = os.readlink(srcname)
                   os.symlink(linkto, dstname)
               elif os.path.isdir(srcname):
                   if self.filterDir(srcname):
                       self.copytree(srcname, dstname, symlinks)
               else:
                   if self.filter(srcname, dstname):
                       shutil.copy2(srcname, dstname)
                       print srcname
                       count += 1
           except (IOError, os.error), why:
               print "Can't copy %s to %s: %s"%('srcname', 'dstname', str(why))
       self.count += count

class LatestTreeCopier(TreeCopier):
   def __init__(self):
       self.count = 0
       self.excludePattern = re.compile(
               '(^\.svn.*|.*\.swp$|.*\.bak$|.*~$|.*\.swo$)',re.I)
       self.excludeDirPattern = re.compile('^\.svn.*',re.I)

   def filter(self, srcname, dstname):
       return (not self.excludePattern.match(os.path.basename(srcname)))
               and (os.path.getmtime(srcname) > os.path.getmtime(dstname))

   def filterDir(self, srcname):
       return (not self.excludeDirPattern.match(os.path.basename(srcname)))


if __name__ == "__main__":
   if (len(sys.argv)) >= 2:
       methods = sys.argv[1:]
       for m in methods:
           method = m+'()'
           print "launching %s" % method
           eval(method)
   else:
       deploy()
       reload()

Java development without Ant

Tue, 12 Jul 2005 16:16:00 +0000

When Ant came out, I welcomed it. It was better than Makefiles for Java, it solved the path and classpath format difference between Windows and Linux, did not rely on shell programs. It did provide a platform independent way for building Java software. At that time, there was another alternative, JMK which had a lisp like syntax, and solved the same problems as Ant. But Ant was backed by apache, and this were the times of XML everywhere. I got more and more fed up with ant scripts over the years. During a long time you had to resort to an XML trick just to include other ant file, you did not have conditionals, but more importantly, it is quite a pain to write long ant scripts in XML. XML is not that good to write logic with, XSLT did not become as big as it could have been partly for this reason.