Chaotic Java

A week ago, Michael Barker wrote a use case for yielder, where he uses the yielding ability to implement “Mini-Axon”, the Kamaelia learning experience usually done in Python, where generators are a built-in feature of the language. I thought it was good to mention it here, to show how yielder can be used for more than just an easy way to implement iterators - for example, to yield results of processing requests as soon as they arrive at the processing box, in the case of Mini-Axon.

Also, if you have downloaded Yielder in the past, go and download the new version which sports two new main features, one of allowing hierarchies under the Yielder class (before, your implementation had to be a direct subclass) and the second is not using debugging information at all, making the tool work with any environment as it relies solely on the bytecode itself. (If you never downloaded yielder, what are you waiting for?)

It’s definitely been an exciting couple of weeks. Ever since I published the code for the yield-like feature (project site here) there has been a few mentions of it over other blogs (Tech Sweep, Neal Gafter’s blog, Lambda the Ultimate and even a reminder of when not to use such a framework from WarpedJavaGuy).

In a very short time, downloads reached 75. And it makes me wonder: How many of these used it, tried it, had trouble with it? I didn’t get any feedback; and I’d really like some! As for myself, I wrote the yielder for tree iterations, and I use it for that at the moment. It works great, creates the implicit stack and returns the elements at the order chosen - and I’m really happy with it. But what about you?

I guess I completely missed out on this because it seems like there was a better looking documentation of Java out there (still a prototype, but is stable) and I kept using the old JavaDoc one.

Anyway, if you’re using JavaDoc and want to have killer web documentations, take a look at this open source project, DocWeb!

I can see one good thing about it and one bad thing: It allows for dead-easy translation of the documentation using the community as translators - Great for open source projects that are relying on the community for so many things, just takes their minds off needing to manage it. However, it is being installed as a servlet which requires you to have some sort of container for it.

I wonder if the open-source hosts (SourceForge, Tigris, Google Projects, etc) are going to start supporting this feature?