"TC" == The Cunctator cunctator@kband.com writes:
Jakob> Semantically there is no big difference Jakob> between dividing an article in subtopics and creating a Jakob> category with a couple of articles in it.
Me> Not true. From a graphical level, part-whole implies a tree Me> structure -- each child article can be part of one and only Me> one parent article. Category-member semantics are more Me> twisted -- articles can be members of multiple categories.
TC> It's true that part-whole implies one-parent relationships.
TC> But Jakob didn't say "part-whole".
TC> So you shouldn't say what he wrote is not true.
OK. I took "dividing an article into subtopics" as "dividing an article into parts". I think there's ways that "subtopics" could be parts, and ways that "subtopics" could be members of a category.
I'd say there's a part-whole relationship if:
* There's an implied ordering of parts. For example, [[US history]] might have parts like [[US history before 1604]], [[US history 1604-1775]], [[US history 1776-1783]], etc.
Contrast [[History of jazz in the US]], [[History of women in the US]], [[History of baseball in the US]]. These might all be members of a _category_ [[US history]], but there's not the same kind of implied ordering.
* There's the one-parent tree structure as mentioned before.
I think we can do interesting things with part-whole relationships. Automated Next, Prev, and Up links might be useful, and an option to read or download or print the whole shebang (container article plus parts) would be neat, too.
Anyways, I've been thinking about parts, as you can tell.
~ESP