Scott McKay reports that Functional Objects wants to open source their Linux Dylan implementation.

I was really excited about Dylan when I was a graduate student in the 1990's. I almost left Taligent to take a job at Harlequin (Functional Objects includes people from the Harlequin Dylan team) working on an OODB for Dylan. The Functional Objects Dylan is the product of some of the best Dylan / Lisp hackers in the world. From a technological perspective this is awesome. Unfortunately, Dylan is now saddled with the baggage of being "that language that Apple abandoned". In recent months Paul Graham's Arc has generated an enormous amount of interest from Lisp (and other people), including myself. I've felt that Arc was to be Lisp's final hurrah. But Arc is nowhere in sight. So where does that leave Dylan? Functional Objects could give the world an amazing piece of technology. I want a better language to work with. Java is fine and C# is basically equivalent to Java. I miss the power of Lisp -- working with Python for pyblosxom is reminding me of that.

Andy Oliver writes.

Ted, I was unaware of your work as I bet was Nick and the Mavenites. However, while I speak for only myself, I bet Nick will agree as will Nicola Ken. We want to write as LITTLE of this as possible and combine as much. If you have something that mostly does it, lets work together because the only thing better than writing code is not having to! I can't really make hide or hair out of it ATM.

I've tried not to be too loud about my stuff at the ASF, because we already had JJAR and Maven, and I didn't want to talk about something that didn't exist. I talked to Kurt Schrader about the Maven stuff at ApacheCon, but got scared off by the size of the Maven code base. The downside of this is that I'm not really as up on Maven as I should be -- I'll be starting on that now. I think that its important to have dependencies both at the build system level and (somehow) at the jar level. That way systems can just use information in the jars to get what they need.

As one example of things that I'd like to enable (someday): If I go into Eclipse and check out a project from CVS, Eclipse should be able to figure out which jar files are needed for the project (aside from checking them into CVS), bring them down into the repository, and set the Eclipse build path. I spent a lot of time trying to track down all the jars for Maven when I tried to get it into Eclipse.

I'll try to fix up the docs and distribution so it's clearer what's going on.

After two weeks of travelling it's good to be home. I arrived just in time to get the new Eclipse M5 build. This time the install was easy -- with M4 I couldn't run the Sysdeo Tomcat plugin until I started with a fresh repository.

The latest version of jardeps includes changes to make the cache pluggable.

In other jardeps news, it looks like there's going to be some work in this area at Jakarta. There is a proposal to make Maven a top level project, and the possibility that the respository code will be split out. There's also some agitation for a new project, Ruper. It seems likely that I'll be teaming up with one or both of these projects -- we really only need one system for doing this. I hope that we will end up with one (not two). I'm going to keep working on jardeps until the dust settles, but I'm also going to start looking at how the Maven repository works.