Axel Boldt wrote:
Our customer here was only interested in math articles, and a download of the whole Wikipedia probably wouldn't even have been attractive for him. So a web site with a static version of Wikipedia that is open to programs like Webstripper would also be a good thing.
Actually, I'm sure that once we offer ready-made HTML trees for download, someone somewhere will set up such a site.
This brings back the issue of classification and categorization schemes (which I think Magnus was working on). Of course, I get gunshy when people turn that into a debate about censorship. Certainly the tool that such schemes give us can also be used to effect censorship. Similarly those who oppose gun control will say that guns have more uses than just shooting people.
I very much support a broadly inclusionist Wikipedia with as few rules as possible for what it allows. At the same time I support multiple classification schemes, and am not troubled at all by the notion that someone might develop a highly censored subset of Wikipedia. If someone wants to set up a subset that would exclude all articles that have the word "and" it would seem goofy, who am I to complain as long as it does not compromise the inclusionist nature of the overall project?
In the past I've tended to support one classification scheme, but really there's no reason not to have multiple classification schemes. All we would need is a way to distinguish which scheme is being used.
Ec