Related link: http://reddit.com/blog/2005/12/on-lisp.html
After a week’s worth of rewriting, reddit now uses Python. Cue storms and furies.
While the explanation given by the reddit developers is that they couldn’t find a good solid open source Lisp implementation that ran on their development and deployment platforms and supported all of the features they wanted, some Lisp advocates accused them of selling out to Venture Capitalists (why not J2EE then?), of being bad programmers (as if something had changed in a week), or of insufficient technical aptitude to do something as simple as resolving to do all development in a different operating system running an emulator for a foreign architecture or deploying to an untested commercial implementation for which the vendor expects ongoing royalties if the project succeeds.
Meanwhile, the developers now have an equivalent site written in a language for which there exists a single high-quality implementation across multiple platforms, for which there exist multiple libraries with documentation, and for which the user community online won’t rush to sacrifice you to a volcano if you don’t toe the dubious and ill-defined line of language purity.
(I suspect that anyone who says “Python is a lot like Lisp” in all apparent seriousness knows very little about either.)
The technical problems for Lisp in the case of reddit are solvable. Some of the posters in comp.lang.lisp realize that the lack of libraries that work well across multiple implementations holds back Lisp… as does the lack of complete features across multiple implementations and multiple platforms.
Still, I suspect they will go unsolved as long as the angry vocal online Lisp advocate community chases away potential allies. (See talk #10 from “Twelve Views of Mark Jason Dominus” — Why Lisp Will Never Win.) I almost wonder if half of the comments in the “on lisp” article came from people who want to see Lisp fail.
John McCarthy, save us from your followers!