[Wikipedia-l] Categories indexed for largest wikipedias

[Erik Zachte]

After rereading my explanation about the article counts, it still seems a
bit vague.
Let me try again:

One a first pass through the database all categories and their mutual
relations are collected.

On a second pass a list is compiled for each article of all categories that
are named in the article and of their supercategories, duplicates are
removed. Then all categories in the list have their counter incremented by
one.

Erik Zachte