Wednesday, October 15, 2014

Next step selecting an alternate programming language

Ok it's down to ADA or Object Pascal (the Free Pascal compiler). I still have to check the license for FP in more details but it looks good. Two languages that have compilers on several platform and libraries to go along. I have worked with Turbo Pascal and Delphi years ago and the Free Pascal version of the language is a very close clone of Delphi so the learning curve should not be too bad. I have played a little with ADA and so far I think the language has very interesting features. I might actually start a more detailed thread about my ADA experimentations. What do you think ? Does anyone out there have any comments about either language ?

Tuesday, August 19, 2014

Separating the user interface form the rest of the code

Before I go back to my exploration of alternate programming languages a short one on design (or is it management ?). Every programmer knows that when you write code that might eventually get used in a GUI application a basic good practice is to always clearly separate the GUI code from the code implementing the actual functions. No big news here. There are a lot of tricks and design principles to help towards the goal of keeping things separate and of course any decent programmer should know about design patterns like MVC. Unfortunately, good design principles can be ignored or neglected if someone else is writing the actual code and additional control mechanism are needed (even you might benefit from additional safeguards). Code reviews can help. Automated unit test will also often help by encouraging developers to encapsulate and write more focused classes. Today's design or management trick is:
- Ask the programmer to write the functions in a separate module (module A)
- Ask him to provide a command line utility (module B) to call the functions in module A
Of course this does not provide an absolute protection against bad programming but it adds another level of control. This can be similar to what automated unit tests provide but not exactly the same since:
- It is fairly easy to check that module B uses A but not the reverse (no circular dependencies). You might not have the same level of isolation for the unit test code and the tested code.
- It is more focused on separating the two parts of the code. The part that implements the user interface (command line for this part of the project) from the part that implements the functions.
= The end product will often be a useful deliverable (this is the best application of this trick)
Of course writing a command line utility should not remove the requirements for automated unit tests.
If the result is not perfect, you or the other programmer doing the work will get another shot at cleaning up the API of module A when the time comes to use the module with the GUI. What do you think about this. Do you have any tricks to help you use good design principle ?

Sunday, August 17, 2014

Escape from the land of C (prologue)

Ok. I'm having a look at ADA. One of rare languages that actually was compared to the aberrant C language to compare productivity. More in my next blog.

Tuesday, July 29, 2014

Is Python a possible alternative for Java

One of the language I evaluated as an alternative to Java of course was Python despite my hatred for languages that have significant white spaces. I did not get very far. In general, the syntax of the language is ok (except for whitespace) but the problem is speed. CPython is terribly slow as in unacceptably slow so I looked at PyPy since this uses a JIT compiler. No luck. It is better that CPython but it is not a substitute for it. You simply cannot assume that the program you ran in CPython will run with PyPy. Some libraries do not run with the JIT and there are a few other incompatibilities. So, Python will not be the successor for Java in my book. What would be your choice of multi-platform open source programming language (really open source - no license for embedded systems) ?

Monday, July 21, 2014

Finding an alternative for Java

Some people at work question the use of Java as our main programming language. Since I am not a Java language fanatic I welcome the discussion however, several of the arguments mentioned against Java simply do not resist close scrutiny. Speed when compared to C/C++ ? JIT compiled Java speed is very close to C/C++ and the problems we have encountered in the past where speed was a problem were mostly related to I/O or ... bad design. Verbosity when compared to other languages ? Not such a big problem since the IDE handles most of the scaffolding (the infamous curly brackets, etc ...) and Java 8 removes some of the verbosity. Certainly not worth switching language specially when you consider the advantage of Java. Advantages that include: an incredible ecosystem of tools an libraries, a very good specification, several of the advantages of modern languages such as Garbage Collection and good support for multi-core processors. In fact, since I love programming languages I did try to find a good alternative because of the license issue. Having to pay for a license on embedded systems is a problem when trying to cut cost. Finding a viable lower cost alternative is not obvious. More on this in the next blog.

Wednesday, December 25, 2013

Another reason to use Netbeans instead of Eclipse

A couple of weeks ago I stumbled on a post by another Eclipse fanatic. The insect ... oups, blogger, was making fun of people using Netbeans. Well, fortunatly for him there was no comments or reply button on the page because I would have replied and my comment would have been difficult for him to handle:
You might like the chaotic and always in beta look and feel of Eclipse however the biggest problem with Eclipse is its buggy Java compiler.
Some people don't even know this but Eclipse has his own "internal" Java compiler. Netbeans on the other hand uses the JDK supplied compiler. Eclipse's compiler has one advantage over the one in the JDK and that is the granularity of the compiled elements. Eclipse will compile and "link" smaller pieces of code. In a few cases where you have to stop debugging and restart in Netbeans, Eclipse will simply recompile the small bit of modified code and keep the debugging session alive. The problem however is that the last time I looked there was still a bug in the compiler forcing us to add useless class cast in generic code. To that you might respond that it is not such a big issue. My response to this would be: well one of the benefit of generics is to remove the need for a lot of type casting in the code so for me it is important. My other response to this would be that it is not the only problem with the Eclipse Java compiler. The reason I switch to Netbeans despite Eclipse being the standard where I work is that for a while in the transition to Java 1.7 the Eclipse compiler would simply not compile my project. This was eventually fixed (a week later) but I simply could not wait a whole week without a working IDE so I switched to Netbeans. I have no problems with that because I have always preferred Netbeans to Eclipse. Aside from what I would call the objective argument (an IDE should compile the code correctly) there is of course a subjective side to my preference for Netbeans. I prefer the more static placement of the Netbeans windows and views. It is flexible enough for me and without feeling chaotic like Eclipse. Also, Netbeans was in front of Eclipse supporting the preview version of Java 1.8 and I would not want to be the guinea pig for the 1.8 internal Eclipse compiler.

Sunday, June 2, 2013

The answer to the question at the end of the previous post

What is the answer to the question at the end of the previous post:
Is using an enum for analyzer family a good way to customize the processing in the client application (CT)
Well the answer in the case of our projet was no. The stable part of the application ecosystem was the CT and the analysis cycle step in it. Adding an enum that we would then use to skip step in the processing would have made it impossible for us to add new Analyzer class without opening the code of the CT to support the new class.(remember the OCP or Open Close Principle)
What is an API designer to do in such a case ? Well, what we did was define an interface for the steps in the analysis cycle of the CT and implement a kind of Decorator for that in all SM(I) instances. Of course, concrete instances of this interface are provided by the IInstrumentDefinition factory. The main interface for this is something like:

interface IAnalysisSteps
void doStep (String stepName);

The CT supplies a default implementation of this that is in fact a supertype that looks like this:

interface IAnalysisStepHandler extends IAnalysisStep
void skipStep (String stepName);

The IInstrumentDefinition interface has a factory method that looks like:

IAnalysisStep getAnalysisStep (IAnalysisStepHandler analysisStepHandler);

The SM(I) implements a version of IAnalysisSteps that will do one of two thing when it gets called:

- Simply call the default implementation of the IAnalysisStepHandler
- Call the skipStep(String) version of the method in the IAnalysisStepHandler

Now SM(I) instances can customize the steps in the CT without actually having to open the code of the CT. What if new analysis steps are added to the CT ? Well, in that case you don't have any choice but to open the CT and to support that you simply have to implement the SM(I) IAnalysisSteps decorator such that if an analysis step is unknown it is skipped. Since this step was originally not present in the CT skipping it should be a good default behavior for old SM(I) instances. I left out some details above like having different parameters for different analysis steps. That however is easily handled using a single additional parameter of type Map<INamedParameter>. If you don't remember what a INamedParameter is simply look at the previous blog entry.