On 6/21/06, Ken McDonald ken@pixologic.com wrote:
The problem with Catgories is that they are "coarse-grained"--it takes some effort to set up a category, so people don't want to create to many of them,
I'm not sure how much easier marking up index entries is. Sure categories take a bit of effort, but anything worthwhile does.
plus there is no such thing as "subcategories",
Well, in fact there are. a category can be in one or more categories. Some find the UI for this a bit subtle, I know I did when I first started using mediawiki. For example, let's say you have categories like Chemistry and Biology. If you put [[Category:Science]] on those category pages they become subcategories of science. Likewise, Organic Chemistry, Physical Chemistry etc. can be subcategories of chemistry in a similar fashion.
So the category machinery in mediawiki isn't as simplistic as it first appears. Wikipedia makes use of it in various ways, the policies for that use have gotten worked out by various Wikipedians who supply the labor to fashion and evolve the various approaches to categorization used there:
http://en.wikipedia.org/wiki/Category:Wikipedia_categorization
plus categories cannot point "within" a page, which is quite problematic for long pages.
As an example of just how nice and precise true indices can be, here is just _part_ of a _single_ index entry in my copy of "SQL in a Nutshell":
indexes ascending creating (PostgeSQL), 59 vs. descending (Oracle), 57 BITMAP (Oracle), 57 ...and so on...
Of course, whether or not such entries are useful depends on the judgment of the person making them. But they certainly can be very useful.
I'm not saying that such an idea MIGHT be useful for mediawiki, but there are alternatives such as the search features of mediawiki to find articles followed by the 'find in page' function found in most browsers. In some ways these can be more useful since they allow a user to approach finding information on his own terms rather than relying on the indexing choices made by an author or editor, this from someone who is often frustrated when a book's index fails to index the terms I'm looking for and I find myself scratching my head for alternatives which might be there.
As an aside, I find myself wondering how one would handle the case of an index term occuring multiple times within a long article. This seems like a case where using the browser's 'find in page' function would be much more natural than hyperlinking back and forth between the index and the article to find all of the occurences.