SKOS
From eg
SKOS is another attempt to define a strong ontology based on technocratic concepts. It has roughly the same problems as SIOC, FOAF or DOAP but remains poorly integrated with those. There is however at least one serious attempt to combine SKOS and SIOC].
Contents |
[edit] not wiki-centric
Like most web-thread-era efforst, it fails to be wiki-centric. Since wiki-based services like Wikipedia are already the most heavily used sources in the world, this could be considered stupid.
[edit] over-focused on blog metaphor
"Most comments on blogs or newsgroups or mailing lists are quite forgettable, and often wrong. They are worthless media for any kind of knowledge management, but wiki (being itself based on software repositories) is more or less ideal. So the concept of an "article" should be the wiki concept of an article, and an "edit" to that article should be the most carefully studied and defined type of "post", and the idea of "discussion pages" (mediawiki) or web-threaded comment threads alongside (tikiwiki) needs to be better integrated" - Craig Hubley, criticizing an SKOS+SIOC integration, who also says:
"Blogs are likewise interesting only because they are RSS feeds so you should put RSS feeds (which are also available automatically from mediawiki to track page-"related changes" or overall "recent changes", or even particular user contributions to a given wiki) at the centre, not the "weblog" as such. A "trackback" is just an unfortunate term for a stable URI such as major public wikis (like WIkipedia) already enforce... a stable URI is meritous in many other applications than blogs, and seems to have been the key to Wikipedia's success. There were always better wikis for any given subject, but it was Wikipedia's simple standard URI "en.wikipedia.org/wiki/...NAME..." that put it at the core of people's reference habits."
[edit] deals poorly with mail/list/group crossovers
"Mail is also increasingly integrated with other web-based services (witness gmail, yahoogroups) so you may have to draw some explicit line between private, shared, and public discourse. Newsgroups are usually wholly public, mail usually assumed private, but there are exceptions, so deal with this dimension somehow in a way that isn't tied to the media itself."
[edit] deals poorly with forward/repost/refactor
"I think you need also to distinguish between forwarding, "quoting" or otherwise just reposting something, vs. "citing" it as a credible "source" (as one might do for a newspaper article or firsthand report). You might want to look at the issue/position/argument form, which is easy to find references on."
[edit] deals poorly with chat and other dialogue mapping
"Finally you don't deal with chat (IRC, "IM", jabber, etc.) at all nor the concept of logs for teleconferences or other phone logs. I think this is a mistake as these media are getting important. In the long run they will often be attached to specific (internal or external) wiki articles, being debates about them, so I suggest you deal with them as a subclass of conversations about pages if you want to remain content-focused in your ontology." See also dialogue mapping and the Compendium tool.
