"Chris Wood" standsongrace@hotmail.com wrote in message news:ch65tl$ose$1@sea.gmane.org...
What I'm arguing is, "is-a" is not going to work, it's too POV. What we have now is sets of related articles, which is all we need (is-a relationships *can* be represented this way, but are more flexible).
Is-a relations work just fine for noncontroversial categories. Quartz is a type of mineral. Bill Clinton is (was) a U.S. President. Arnold Schwartzennegger is a governor of California. Canada is a country in North America. Then there are some areas that are a little more controversial, due to judgements about the classification system used: should a particular author be included in the category Fantasy writers, Science fiction writers, or Horror writers, or all three? But I think most such questions could be decided relatively amicably. (I realize that there are problems even with such simple classes as well--e.g., the discussions about which counties should be included in the Europe template--the line dividing Europe and Asia is not self-evident at all points).
Where is gets much more difficult is where there are intense POV values built into the categorization (especially where the defining criteria depend on subjective determinations). Such as whether something is a work of propaganda. Or whether a person is an alcoholic. Or if an incident was an act of terrorism or of patriotism.
But I do not see how your distinction between related-to and is-a makes any difference in such situations. I think the difficulties of NPOV categorization are not that different from that of NPOV language in an article in general. The language used to describe a person/thing/event inevitably carries some subjective value judgements about the topic. The criticism you make of categorization seems to be a variaton on a critique of language in general. In your earlier post, you
I guess I simply do not understand how related-to is any better than or so very different from is-a categorization. If you are suggesting that we should have a very flat category structure with many more categories for each article, then I very strongly disagree. Hierachichical categorization is useful, despite the inherent difficulties with any categorization schema. I do not see any significant attempts to make categorization in Wikipedia is an endorsement of any single ontological worldview--there are mutliple competing/complementary hierarchies. I think your assertion (in your earlier post) that "Sets are much less POV" is simply wrong. A set is just as POV as a hierarchy. Whether there is a list/set of alcoholics or of propaganda or of acts of terrorism--inclusion in such a list/set is expressing a POV just as much as if including within a hierarchical categorization schema.
Bkonrad