The Library Problem
Bill de hÃra's "Better is Better" got me thinking about the library problem for languages. The problem for languages with a number of implementations is that the possibility for . On the other hand, the Perl, Python, and Ruby communities don't have to worry about language portability, only platform portability, and I think that's contributed to the large set of libraries available for these languages.
The quote that really got me from Bill's post was "Thus, while the book, ANSI Common Lisp, does a great job telling you how to on generate HTML documents, it doesn't tell you how to serve them up or read them from disk or render them.". This is true, Paul spends minimal time on this. However, I think that this is the role for a Cookbook-style book. There is work on a cookbook for PLT Scheme going on in the Schematics project right now, but the Common Lisp Cookbook is pretty thin.
Bill also points out that the prevailing attitude in the Lisp community is DIY, and that's simply not a viable strategy these days. For example, I've wanted an SHA-1 implementation for Scheme, but the best option right now seems to be to write my own, and I just don't have enough need or desire to do it. Bill mentions that there's no standard library for things like database access and GUI in Common Lisp. This is true, but I can say for a fact that Franz has some support for these things in Allegro. But this means vendor lock-in at a much lower level than would be ideal.
This brings up a second problem: Common Lisp has a second problem in that there's a number of commercial implementations; in fact, the only really worthwhile implementations are commercial. This puts the onus onto the company to provide the libraries. Java, VB, MFC, and C# have the backing of big companies who can throw a lot of dollars at the problem, but small companies have to choose their targets more selectively. But I have a feeling that people who get something for free are more inclined to contribute libraries back into the language, while people who pay feel more like their vendor owes them the libraries, and anything developed by the customer is proprietary. I don't have hard evidence for that, but that's my suspicion.
That said, I still feel like learning Lisp has been very valuable. The syntax gets to be addicting; Python and Ruby may be more familiar, but the natural recursiveness of Lisp has an appeal all its own.