From: Erik Moeller on Thursday, December 11, 2003 5:52 PM To: wikitech-l@wikipedia.org Subject: Re: [Wikitech-l] Categories simplified

Cunc-

...

Why do we need [[Category:Mathematics]] and [[Category:Biology]]

when we

...

already have [[List of xy topics]]?

1. Adding metadata to the article it belongs to makes it easier to

find

and update. Having it elsewhere would in fact add aditional workload

and

require additional learning, because to verify whether a

categorization is

correct, people would now also have to check all the [[List of xy articles]] page where it could be or has been entered. With a

Category:

approach, they only need to keep an eye on a single page (the

article),

and the categories themselves are structured through subcategories and therefore easy to pick.

2. Category pages are not articles. Like talk pages and meta pages,

they

should be logically separated from articles, which has numerous

benefits

(easier searching/filtering, counting etc.)

If we accept this principle, we should replace all of the [[list of ]] entries with [[Category:]] entries.

I'm not quite entirely convinced that the healthiest thing to do is to separate category/list pages from the main namespace, but it's not a bizarre concept.

3. In terms of information organization, your approach leads to

bloated,

big pages (the list of lists is going to be hundreds of pages long,

the

lists themselves are going to be 30K and more), whereas the category approach generates all long pages automatically from the information stored inside the articles. It is much easier to handle.

That doesn't seem like a great argument--the list of categories is going to be big and bloated.

Rather than looking for points of conflict, I think the more useful approach is to try to figure out what our specifications are.

For example, the hackish "list of" approach (which could have been our "category of" approach--the name is immaterial) has plenty of problems which we would want an alternative implementation to eliminate, for example the basic problem of not being in sync with the entries in that category. If someone makes an entry about a random American actor, he currently has to check if that actor is listed on the list/category page.

However, one advantage of the "list of" approach is that it allows people to make more interesting groupings within the list than simple alphabetization.

So what I'd like to see (and maybe it's on some meta page I've failed to check--again the problem with having this discussion not be directly linkable to Wikipedia content) is what it would look like for some example entries.

Right now, for example, the Kevin Bacon "What links here" looks like:

* 2003 in film * A Few Good Men * Apollo 13 (movie) * Bacon number * Cigar Aficionado * JFK (movie) * July 8 * List of famous Philadelphians * List of male movie actors * List of people by name: Ba-Bd * Mister Roberts * National Lampoon's Animal House * Six Degrees of Kevin Bacon * Small world phenomenon * Tony Banks * William Rufus Shafter

That's a pretty nice set. It's not perfect but ideally I'd like to see us figure out ways to improve upon "What links here" rather than ignore it...