Paul [the interviewer]:
06:05 OK, sounds good. Going a little bit broader than those questions then, back in 2001 in that Scientific American article, you and the other authors painted a very broad grand vision of where the Semantic Web could take us. Did you think we'd be closer to that seven years on?
Tim: Well, for one thing that article was, I think, too sci-fi. I think, that really what we have... the message has been... it was looking too far into the future. It imagined the Semantic Web was deployed, and then people had made all kinds of fairly AI-like systems which run on top of that.
(I'm the author of the post, but I didn't post it here - thanks to whoever did)
I've got no problem with XFN and/or FOAF - I've built sites that support them. The problem I see is that advocates for those formats concentrate on the easy stuff - a data format - while ignoring the hard stuff - spam, granular privacy controls.
These problems DO need solving now, though - without them these technologies are about as useful as meta-data in webpages. Maybe they should relaunch AltaVista circa 1998 as a new "Semantic Web" search engine..
The Norvig quote is very misleading. I thought he was both forthright and realistic about what SW and sw mean today, and to Google especially, during the Q&A of that talk.
I disagree quite strongly that it was misleading. That was a direct quote, and it appeared to me he was making fun of the person who got up and stated "I see myself as a Semantic Web evangalist... what is Google doing about it." Norvig went on to say that Google was always watching what was happening in the space.
I don't get spam in my feed reader because I select the feeds it's reading. Does RSS "cope with this thing we call the web"?
It would be great if I didn't have to select the feeds, if my feed reader would crawl the Web and show me things I like without forcing me to wade through heaps of crap. I don't think that's very realistic though, so it's still worth my effort to choose trusted data sources myself.
The same is true for many applications of the Semantic Web. If you want to buy a house, you won't send out a crawler looking for listings matching your criteria; you'll go to a site that aggregates real estate listings in your area. This isn't so different from how things work today.
Spam is a problem and fancy trust stuff is absolutely needed, but RDF & friends have a lot of value even without fancy trust stuff. The problem is (and has always been) convincing people with interesting data that they should publish it.
Why are you comparing RDF with syndication formats? Atom and RSS are formats for data that fits a specific pattern, RDF is a generic data structure. A better comparison would be RDF and XML.
If there is some advantage to putting (e.g.) real estate data inside an Atom or RSS feed, you still need a format to represent the data itself. RDF's advantage is that it's extensible and makes it very easy to combine independent datasets.
I don't think that Atom/RSS/XML are easier to use than RDF; they're just more familiar.
Seems like the argument is that the Semantic Web (or downcase) has merit but requires coded trust relationships between data providers to work. So, the problem is that intelligent people are busy working on things that rely on the problem already being solved--which it very much isn't. That's a cogent argument.
So I've recently been dabbling in XFN and FOAF, which are like baby steps towards some kind of coded trust relationship and the problems there are: 1) they are insufficient for non-personal relationships, 2) the cost/benefit for personal users maintaining these coded relationships isn't there and 3) no one appears to be solving these problems at the moment because they don't really need to be solved right now.
Nice article (I like your style)... it would be better if people referred to semantic web technologies as technologies and facilitators of other applications, not as the abstract, "future" concepts. I address many of these on my blog -- you may find some of my MI topics interesting/applicable.
http://talis-podcasts.s3.amazonaws.com/twt20080207_TimBL.htm...
And particularly this passage:
Paul [the interviewer]: 06:05 OK, sounds good. Going a little bit broader than those questions then, back in 2001 in that Scientific American article, you and the other authors painted a very broad grand vision of where the Semantic Web could take us. Did you think we'd be closer to that seven years on?
Tim: Well, for one thing that article was, I think, too sci-fi. I think, that really what we have... the message has been... it was looking too far into the future. It imagined the Semantic Web was deployed, and then people had made all kinds of fairly AI-like systems which run on top of that.