- From: Danny Ayers <danny.ayers@gmail.com>
- Date: Sat, 4 Aug 2007 18:05:57 +0100
- To: kuldar@csse.unimelb.edu.au
- Cc: "Valentin Zacharias" <Zacharias@fzi.de>, "Pat Hayes" <phayes@ihmc.us>, "John F. Sowa" <sowa@bestweb.net>, "[ontolog-forum]" <ontolog-forum@ontolog.cim3.net>, "Ivan Herman" <ivan@w3.org>, "Juan Sequeda" <juanfederico@gmail.com>, "SW-forum list" <semantic-web@w3.org>
On 04/08/07, Kuldar Taveter <kuldar@csse.unimelb.edu.au> wrote: > I must say that I totally agree with Valentin. It is not possible to solve > the interoperability problem by imposing > a set of standards "from above". The world is too diverse a place to assume > that everyone is going to use them. Hmm, I generally agree with the "from above" point, yet I find myself disagreeing quite strongly with Valentin... It does seem to be reasonable to try and build a global data system based on what we've already got with the Web. So what might be the simplest and most flexible approach? The Web essentially consists of documents and links (enabled by HTML and HTTP). A link is a relationship between the current document and another document. There's a familiar data model here, if you generalise from documents to arbitrary entities, and give the relationships types. This is very much in reach of contemporary software practice: entity-relationship style data has echoes in object-oriented programming and can be mapped to Codd style relations. Conveniently it also maps to some of the logical predicates of FOL. Documents are named uniformly with URIs, so it makes sense to name the entities and relationships of this extended Web following the same system. One advantage of the Web over other approaches to hypermedia is that it takes into consideration the broken link, the missing document - the 404 Not Found is very much a feature, not a bug. In the data model scenario, there's a reasonable analogy with the open world assumption. Allow for unknowns. ...and there in a nutshell you have most of RDF, the core of the Semantic Web technologies. (Ok, a lot of logician-hours and practice went into producing the current specifications which allow this naive post hoc derivation, but I reckon it's still valid). There may be arguments what's best to layer on top of this system, and it's interesting to note that certain aspects of RDF that go beyond these basic components (bnodes, RDF reification) are actively avoided by some (er, increasingly many) implementors. But this foundation is very much capable of supporting a whole range of alternatives whilst maintaining Web-style compatibility. It's already demonstrating utility, based on tools ranging from data(/abstract syntax)-oriented query setups to more formal DL-oriented reasoning systems. Re. "The world is too diverse a place to assume that everyone is going to use them." What of HTML and HTTP? Cheers, Danny. -- http://dannyayers.com
Received on Saturday, 4 August 2007 17:06:02 UTC