Carlos found a follow on paper to the Norvig paper.  It's interesting to note that the author, Greg Sullivan, used to work on Dylan and is part of a research group that has a few ex-Dylan folks in it.

For those following this topic, this comparison of Apache vs Yaws appeared in my Aggie page today. Yaws is a web server written in Erlang, another functional language. The benchmark shows Yaws scaling up way beyond Apache. Oh, and that's Apache 2.0 (sorry httpd guys)

Ok, so I grabbed Stuart Langridge's "Nice Titles". We'll see if it's not too gaudy.

Adding nice titles made me realize how many title attributes I'm omitting. I'm editing blog entries by hand in Emacs, and it's kind of a drag. I should try to discipline myself to do this using Mozilla's composer. I just started using 1.3b and it looks like Midas is active, so maybe I should hack pyblosxom to have an input form based on Midas...

Gordon Weakliem writes

Lots of people talking about functional languages these days: Charles Cook, Chris Double, James Robertson. Even Sam Gentile's picking up on generative programming. Something's in the air, for sure.

It's not the air, or the water. There are real deficiencies with the language tools that we are using today. Deficiencies for which solutions have existed for 10-20 years, depending on how you count. Now that the world accepts that "managed" languages like Java and C# are useful, it's only a short jump back to Lisp.

Bill de hÓra tells another classpath related horror story. This time its having multiple incompatible versions of a jar in the classpath. Problems like this are why I think that a solution to the classpath problem has to address both the source level and the binary level. In the discussions on Ruper (which seem to have petered out), people were arguing that we only need to solve the problem at the project specification level (via the Maven POM or the Centipede equivalent). But Bill's problem shows that we need to also solve the problem for binaries. In fact, Bill proposes Jar manifest annotation as part of a solution to this problem. Which is exactly what jardeps does.

Steve Conover seized on Bill entry as an opportunity to write a tool that finds duplicate classes in the classpath. This is great. Now we still need a way / place to get all of this stuff together in one place so that people can deal with classpath problems.