I bought a new logo for the site a while back. I finally got a free moment to use it on something more than a slide show. Check it out both on CGN and CGN Connections

Here is a little self-promotion for you. My new article on using XSLT to get generational portability is up.

Fabian Pascal has published a private conversation with him that I had when I was interested in his opinion on an article that I was writing for DevX. I'm not quite sure what the point of publishing the conversation was.

Here is an interesting article on SourceCafe.

My favorite line from this blog entry about the complexity of J2EE. In regards to J2EE simply printing "Hello World!" on a web page; "Indeed, J2EE is the wrong tool for this job." How can you be the wrong tool to print "Hello World!" The point of "Hello World!" is to provide the simplest possible implementation with the tool or framework.

Anyway, I thought the blog entry was cool overall, I just though that part of it was funny.

I met with Terrence Parr earlier this month. He had some interesting ideas about enforcing the separation between business and presentation logic and has written a new paper (PDF link) on the subject. Worth a read, especially if you are a Java guy.

There is a new interview with Norman Richards up on CGN. Norman wrote XDoclet in Action, which is the sister book to Code Generation in Action. XIA is focused specifically on XDoclet and is a fascinating read.

I gave a talk at the 2003 International Ruby Conference this week. I think the talk went over pretty well. I gave my presentation on the first day and the two days following we talked on and off about CG approaches to various problems.

The other CG approaches that were presented were very interesting. One of the unique attributes of Ruby is how literate the code is. It looks and reads in a very simple and structured form. So people were using Ruby code to build the model in memory and then using generator code to use the in-memory form of the model when generating the code. It looked and worked really nicely. I'll have to think about the advantages and disadvantages of this apporach. Hopefully I will be able to interview some of the other presenters and put those up on the site for some feedback.

One of the more intruiging uses was using a code generator to build an adapter layer between Ruby and Java's Debug Wire Protocol which allows you to connect to a running JVM and debug the code over the wire. It was really cool to see a running Java process was then started and stopped by an external Ruby process.

The author had hand written the JDWP spec as Ruby code which was, in effecft, a model in memory. This model was then traversed by a generator which build Ruby classes on disk to do the marshalling for each of the various binary packet types in JDWP. The binary marshalling stuff was ugly, but it is in every scripting language. I don't think there is an elegant way to do that. The cool thing was that all of the code was documented with the spec itself, so once the code was generated you could use RDoc (Ruby's equivalent of JavaDoc) and create HTML pages which had the entire JDWP specification as well as the Ruby methods to make the calls.

One of the great things about Ruby is it's users. These are some clever, intelligent and resourceful people that come up with extremely elegant solutions to difficult problems.

Alan Fisher of Ironspeed has a new article out on DevX. it's well worth the read.

I've been reading up on SVG. It's really cool. The troubling part is the effects. They are very powerful but the interface is really raw. What I need is a set of recipes for various Photoshop or Fireworks effects, like Inner Glow, Outer Glow, Bevel, Eboss, Drop Shadow, etc.

Anyway, to demonstrate the coolness, here is an interactive SVG map of the US. You will need to get the SVG plugin from Adobe if you don't have it already.

I used to think that with rampant abuse of the term Patters the industry had turned a good idea, a nomenclature for industry accepted software architectures, into a bad idea. But now I look at some of the books I have been reading recently and it's obvious to me that these books, and their associated products would have been much harder to produce if everyone had a different notion of what a dependency pattern or a factory pattern was.