Cobra Popularity

[cmack]

Just found out about Cobra. I think this has the nicest syntax out of any of the statically typed languages I've seen so far.

A few thoughts on things that might increase it's popularity.

Cobra for the JVM, which I've read it's on the way. There are a number of new-ish statically typed languages that are targeting the JVM, but I haven't found any that look as nice as Cobra. Cobra could really stand out here, since most of these languages are variations on Java, perhaps throwing in a bit of functional programming and cleaning up syntax (though most don't go very far here). Perhaps an Eclipse or Netbeans pluggin would also help popularity.

Python is very popular in the scientific computing world because there are so many great numerical/scientific computing packages for it. Packages like Numpy and Scipy have syntax and features that are similar to MATLAB, plus you get all the data structures and packages in Python that you don't normally find in MATLAB-like languages. It may be fairly easy to convert some of those Python packages to Cobra libraries. I could see Cobra taking off in this area. I'd be willing to help build some matrix/numerical libraries in Cobra, if others felt there was a need.

In general I think packages that make programming easy, whether they come standard or separate, are very important. Python has some great ones that just make life easy and programming fun. Python also tends to make optional packages standard after while (if there's a good enough reason and the packages are well designed).

[Charles]

Making Cobra more popular will require a team effort with multiple people contributing different things.

Also, I thought that we could go with 2 releases per year, but it appears that is too few as people question if the project is still active. So I will switch to 3 times per year, or approximately every 4 months.

You mentioned scientific computing which is interesting because in 2003 I attended the SciPy conference where many of the presentations and side conversations were about how to overcome Python performance issues. There were numerous options, many of which I hadn't heard of before, each with its own set of pros and cons. I became more convinced that a good language design should strike a better balance between code-time efficiency and run-time efficiency.

[cmack]

In my line of work, bioinformatics, much of the scripting, prototyping and data analysis is done in Python, but when dealing with larger datasets and/or computationally intense tasks one usually needs to move to another programming language such as Java or C++. I enjoy programming in Python and always dislike having to switch to lower level languages for these computationally intense tasks. This is one of the reasons that Cobra seems so promising. I am tempted to start working on a matrix library for Cobra, along with a Bioinformatics library. It would be great if there were a few other people that also wanted to work on either of these. If libraries like this are available, I think many Python programmers would start using Cobra.

[Charles]

I'll volunteer to do code reviews and provide advice on using Cobra.

[nerdzero]

I can't say I know much about the bioinformatics domain, but if there's any .NET libraries already available those could be used pretty much out of the box with Cobra.

On the other hand, writing (or rewriting) anything in Cobra is really fun

[cmack]

Java has a real nice bioinformatics library, BioJava. I looked for similar C# libraries, but haven't really found anything as extensive. Python has one called Biopython, which I use regularly and could perhaps serve as a template for one in Cobra. The bioinformatics (and more generally, scientific computing) community tends to do a lot of work on Linux and Mac systems, so there may not be as many available packages for .NET languages. Though this is more of a guess on my part, since I'm not too familiar with the .NET world.