- From: Tim Berners-Lee <timbl@w3.org>
- Date: Mon, 28 Feb 2000 10:26:53 -0500
- To: "Bill dehOra" <Wdehora@cromwellmedia.co.uk>, <www-rdf-interest@w3.org>
-----Original Message----- From: Bill dehOra <Wdehora@cromwellmedia.co.uk> To: www-rdf-interest@w3.org <www-rdf-interest@w3.org> Date: Wednesday, February 23, 2000 9:43 AM Subject: RE: A certain difficulty > >:As it happens I think RDF is nowhere near as difficult as people think. > >RDF *language* certainly isn't that difficult. But the *modelling* that is >required to use RDF is, and I think that is the problem these engineers are >having. I think there will be a paradigm shift after which it will appear obvious. As for the web in the first place. >This issue applies also to XML schema and DTD design. You can't just knock >off a schema system, set of standard DTD's, or RDF and expect it to be any >good at the scale of a business enterprise, anymore than you can knock off a >database model at the same level. You need clever specialised people to do >this. Again, by analogy with hypertext, you used not to be able to create a documentation system and have it scale - before the web, which decentralized the model and therefore gave everything you do the ability to fit in. >:And it is incredibly significant. I don't think it's an exaggeration to >:say that RDF and RDFS will become *the* most important XML technologies >:as we try to build a web of information - not presentation. In fact if >:you re-read all the hype and 'promises' of XML, you'll find that XML on >:it's own cannot actually implement them - RDF can. > >And what of certain *very* hard problems with regard to a computer's ability >to handle meaning? RDF does not guarantee shared semantics between computers >any more than speech acts, frame theories or distributed systems have in the >past. No, there *is* something fundamentally different between most distributed OO systems and frame systems and RDF. In a frma system, fundamentally, the designer of an object class defines the set of properties an object may "have". This doesn't scale, as it doesn't allow anyone to say anything about anything: you can only say what the class designer said you could say, and in some syetms you can only say it if you have write access to the object. The RDF model in which properties are essentailly first class objects independent of classes (though constrains can later be expressed) is fundamentally more weblike, and therefore scalable. What we deserately need is some more tools for example for expressing equivalence between separately invented ideas. See http://www.w3.org/DesignIssues/Evolution for my nurdlings on this >RDF is for machines not people. I prefer to say it is for hard data - the sort which machines typically (but not always) process and which humans still find most comfortable with in certain circumstances. [...] >I don't dispute the power of RDF at all for certain tasks. I do dispute the >need for the w3c to invent it all in the first place, other far more >powerful formal systems and logics could have adapted for the job. There was a choice, to go for the minimalist model or a full-blown higher order logic. The first can represent anything by reification and the second by its inherent power. The plan is to build the second on top of the first promptly. It would have lead to a huge food fight to have started with HOL. >-Bill Tim BL
Received on Monday, 28 February 2000 10:26:58 UTC