Hi

I added a comment to Bug 1012 and I thought I'd explain it here little bit more. The problem:

When requesting category one will get the raw Wiki-text as a response which will be like this in the case of http://en.wikipedia.org/wiki/Category:Geography_by_country?action=raw (or edit):

[[Category:Geography|*Country, Grouped by]] [[Category:Categories by country]] {{catAZ}} ... some few lines more...

But what an API or a crawler would expect in this case is the fully evaluated Wiki-text (or XML). This will be the same content of the HTML result page but with no style and as a - potentially large - contiguous list of countries (in Wiki-text or XML again).

I now have looked again what has been proposed until today and I found "Bug 208: API for external access" some requests at Wikitech-l like a "Minimalistic Web-API for use by Tools nad Bots".

All alternative solutions proposed so far, like the Phyton Framework (being a HTML screen scraper) or Perl (reading SQL-Dump) do not solve our problem. So I really want to begin to play around with the Wikimedia-PHP code.

Any comments or help?

Stefan