I've posted a few entries in the past questioning the value of the Semantic Web as currently envisioned by the W3C along with its associated technologies like RDF and OWL. My most recent post about this was On Semantic Integration and XML. It seems I'm not the only XML geek who's been asking the same questions after taking a look at the Semantic Web landscape. Elliotte Rusty Harrold is at WWW2004 and wrote the following opinions of the Semantic Web on Day 4 of WWW2004

This conference is making me think a lot about the semantic web. I'm certainly learning more about the details (RDF, OWL etc.). However, I still don't see the point. For instance what does RDF bring to the party? The basic idea of RDF is that a collection of URIs forms a vocabulary. Different organizations and people define different vocabularies, and the URIs sort out whose name, date, title, etc. property you're using at any given time. Remind you of anything? It reminds me a lot of XML + namespaces. What exactly does RDF bring to the party? OWL (if I understand it) lets you connect different vocabularies. But so does XSLT. I guess the RDF model is a little simpler. It's all just triples, that can be automatically combined with other triples, and thereby inferences can be drawn. Does this actually produce anything useful, though? I don't see the killer app. Theoretically a lot of people are talking about combining RDF and ontologies from mulktiple sources too find knowledge that isn't obvious from any one source. However, no one's actually publishing their RDF. They're all transforming to HTML and publishing that.

I've written variations of the same theme over the past couple of months. It's just hard to point at any practical value that RDF/OWL/etc provide over XML/XSLT/etc for semantic integration.


Sunday, May 23, 2004 2:08:13 AM (GMT Daylight Time, UTC+01:00)
Well, it provides generic inference and validation with proper logical formalisms, URI identifiers to allow graph superposition, and all the advantages of directly representing statements rather than having them implicit in a document tree.

A good comparison of XML and RDF is http://www.w3.org/DesignIssues/RDF-XML.html.

As an example, I can take a hundred documents in OWL about a variety of things, push them straight into RACER or another reasoner, and it will automatically combine those statements into one graph. I can then issue queries of any complexity (e.g. in OWL-QL or description logics), and it will answer INCLUDING ALL OF THE IMPLICIT KNOWLEDGE. E.g. if a 'parent' relationship is the inverse of a 'child' relationship, then for each child the parent is implicitly known, and will bind in queries. I can also describe classes, sets, lists, disjunctions, and everything else that's logically useful, and perform inference and querying with them.

Try getting that with XSLT. Hell, try combining two XML trees and still have it make sense!

"What pieces of writing have Tim Berners-Lee as an author?" is a reasonable query for the Semantic Web, though it would handle equally well "what people who live close to Stratford own a car that has been parked in Reading in the last year?".

XML is a good interchangeable document format, but it's no good at knowledge representation.
Sunday, May 23, 2004 11:26:53 AM (GMT Daylight Time, UTC+01:00)
Heh, what Rich said!

One of the key things in the comparison with XML/XSLT etc for semantic integration is that XML alone has no semantics. The result is that different XML formats/languages have their semantics defined completely independently, so in the general case there is no common ground for integration.

Another key aspect that Rusty misses is that RDF, as well as being describable in triples, is a graph model. This is a better general fit for real-world data than (a collection of) trees. I believe It's also a far better fit for the arbitrarily-connected web. Something that Microsoft appear to have recently discovered.

XML is a very useful document/data syntax, but everything is at the syntax level.

Probably the strongest cases for the use of RDF/OWL are "try it" and "just watch". Play with the stuff and find out for yourself what it's good for. But whatever, people are adopting it and benefit from it, and the impact will be apparent on the web as the islands join up.

Oh yes, and RDF brings what WinFS brings (and more), see:
Sunday, May 23, 2004 6:32:44 PM (GMT Daylight Time, UTC+01:00)
I'm working on interoperability at the moment, and it's an absolutely central point that XML provides interchange (syntax) interoperability, but RDF and OWL provide semantic interoperability. I.e. XML is a data format that gives you a parser toolset, character encoding doodahs, etc.

Anyone who thinks that XML 'does' interop is ignoring the mess of implicit ontology that they're required to transmit.

Furthermore, any effort to "fix" XML so that you can provide explicit ontology is simply re-inventing the Semantic Web.

Danny: As a long time reader of Raw, I never thought I'd have _you_ agree with _me_ :D
Thanks for bolstering my ego!
Monday, May 24, 2004 3:04:40 AM (GMT Daylight Time, UTC+01:00)
You trot out the same old examples the Semnantic Web folks do and I still don't see how I can't use XML/XSLT to solve these problems in a practical manner. For example, there are 3 major syndication formats around today (ATOM, RSS 0.91/2.0 and RSS 1.0). To merge feeds in all three formats into a coherent whole, in RSS Bandit I transform each format into a canonical form (akin to RSS 2.0) and operate on that canonical form. In theory, I could do something similar with RDF. However I fail to see what using RDF buys me over my current approach except that I'd be using a different technology to perform the same transformations and queries I perform with XSLT and XPath today.

That's what I find hard to grok about the Semantic Web family of technologies. Give me practical reasons for changing my approach not philosophical and theoretical reasons why RDF/OWL/Semantic Web.

I read the linked TimBL paper and it left me with more questions than answers. I'll probably blog my opinion of that paper later this week.
Monday, May 24, 2004 5:21:17 PM (GMT Daylight Time, UTC+01:00)
My most recent blog entry contains some of my own thoughts on the different role of RDF versus XML.

Basically, it comes down to RDF for data modelling and XML for surface syntax.

See http://jtauber.com/blog/2004/05/15/xml_infoset_and_xml_schemas_versus_rdf_and_rdf_schemas
Saturday, May 29, 2004 10:04:03 PM (GMT Daylight Time, UTC+01:00)
Dare, as thought exercise, take your normalised RSS and now try incorporating some information about the people that are authoring those feeds - quite a few folks are now using XFN, there's also OPML blogrolls. Now try incorporating some event-based information, like who was at such-and-such a conference, maybe from the XML version of iCal.

How would you get some information like, "what did the people that attended WWW2004 that are known by Rich say in their blogs that week?".

This, and a lot of things like it are conceptually trivial using RDF, and this kind of data is already available - RSS 1.0, FOAF, iCal/RDF. That kind of query can be expressed using RDQL. Ok, you might want to shift to using an XML DB or even a RDBMS, but I doubt whether it'll be possible to plug in whatever specific RDF languages you like and reuse *exactly* the same system for a completely different domain.
Comments are closed.