- Plural when the object being categorised is plural, singular when there is only one such object. At least, that's the sort of plan. --prime mover (talk) 15:11, 28 January 2017 (EST)
How large should categories be?
- The trouble with writing rules for this stuff is that in cases where the rule would not work, there would be a lot of trouble bending the natural structure out of shape in order to make it fit something arbitrary. --prime mover (talk) 17:45, 25 August 2017 (EDT)
- It's all been done by the pragmatic approach that if a category spreads over 2 pages it's too big. In the early days before sub-specialisation of categories (before we'd really learned how best to utilise MediaWiki) I found myself spending ages searching through a large number of proofs in Group Theory trying to find whether a specific basic result had been covered. So "small is best".
- On the other hand, splitting them down too far will give the opposite problem: it will difficult to find the correct category because now they were either nested too deep to make it easy, or there are now more subcategories than can fit on one page (over 200).
- So an attempt is made to find a comfortable middle ground.
- (The gradual move towards putting multi-proof results into their own category may be a step in the right direction, but this itself bulks up the subcategory count, so I have in mind another template-based idea: a template called "ProofCategory" which puts a given proof page into a category ".../Proofs/(page name)" or something, so a category has just one subcategory for "proofs" which would then not bulk up the subcategory count, and simplifies the page development process. Early days yet.)