The following is a discussion from #swig on irc.freenode.org - the Semantic Web Interest Group. It's logged here if you don't believe me: http://chatlogs.planetrdf.com/swig/2008-04-15#T10-32-11. Edited slightly for clarity.
I think the semantic web is an extremely useful tool, but as I mention down the bottom, I probably would have agreed with Francis Bacon that cutting up animals in the name of science was a good thing at the time. For the record, I don't believe this.
Reading the comment first might help.
naught101: my comment down the bottom, would love feedback from anyone here
bengee: simplification is a feature, not really a problem
bengee: URIs and triples reduce the complexity to a level that computers can do useful things with it
bengee: e.g. <#product> :rating "***"; :rating "****"; :rating "**". what might seem contradictory to you may be very useful to an app
naught101: na, I wasn't talking about that kind of information bengee
naught101: Say philosophy for instance... let me find a nice quote
naught101: Only when the last tree has died and the last river been poisoned and the last fish been caught will we realise we cannot eat money.
naught101: * Cree Indian Proverb
naught101: Obviously a computer could use this sentence, but would it be able to use it a a way useful to humans?
naught101: Obviously it's not TRUE, as most of us already know we can't eat money
bengee: <#only1> a :CreeIndianProverb; rdf:value "Only when..." .
bengee: that triple could be useful for programs that list proverbs
naught101: yeah, sure, but that's triplification ABOUT the proverb, not about the information contained within the proverb
bengee: well, then you have to increase the granularity if your app wants to provide richer functionality
naught101: how do you mean?
bengee: extract more triples from the human-readable text
naught101: but what triples could you extract from a one-sentance text that has no quantitative truth, but which holds more qualitative truth than many many paragraphs of, say, a science text book?
bengee: you may misunderstand what the semweb is mainly for
naught101: sure, that's a true sentence :)
bengee: it's not for implementing automated philosophers, or compete with humans with respect to intelligence
bengee: well, OWL folks might disagree with me here ;)
bengee: the more rewarding approach (IMHO) is to think about use cases that semweb tech *can* enable/simplify, not to think hard about things that are near-impossible for computers in general
naught101: I agree. but I'm not talking about what the semantic web should be
naught101: I'm talking about possible problems with what it currently is
naught101: I mean, I don't want bite-sized chunks of information taking over the world of ideas
naught101: I think the philospohy or the public is degraded enough without chopping it into bits even more
bengee: oh, semweb tech can clearly improve the distribution and discovery of ideas
bengee: just like the web did
naught101: but it could also hide them
naught101: I don't think the 'web did, neccesarily
bengee: you just google'd WRT, no?
naught101: I don't think finding an accronym compares to finding meaning in life.
naught101: (if I sound like I'm attacking the semantic web, I'm not, I'm just exploring ideas)
bengee: yeah, don't think I can contribute too much here, sorry.
naught101: no worries :)
naught101: I see it something like baconian/descartian science. it's useful for finding out the little bits of information, but it's not particularly useful for figuring out the interrelationships, or looking at the information holistically
naught101: I mean, for example, the semantic web can take information from a wikipedia article, but it couldn't write a wikipedia article
kjetilkWork: I don't think it is a very significant goal of the semweb to produce that kind of information
kjetilkWork: we have a billion people out there that can do that much better
bengee: the semweb can be a great aid in helping you write the article, though
naught101: sure, but thinking of the possibility the semantic web a large part of the web should probably include thinking about what it can't do, and how to not impede that work
kjetilkWork: rather than the AI world of natural language analysis to reason and infere relation, I think the semweb is much more about using the collective intelligence of all its users, i.e. real intelligence
naught101: bengee: yes, it could. it could also be a hinderance (information overload)
naught101: kjetilkWork: good. I like that. I just hope we're collectively intelligent, and not collectively stupid :)
kjetilkWork: well, that's what it means to me, at least
kjetilkWork: hehe, yeah
kjetilkWork: I think semweb can help us be collectively intelligent rather than stupid, though... :-)
naught101: I haven't got that far yet