- From: Dan Brickley <danbri@w3.org>
- Date: Fri, 23 Aug 2002 17:50:01 -0400 (EDT)
- To: Paul Prescod <paul@prescod.net>
- cc: Dave Beckett <dave.beckett@bristol.ac.uk>, <www-rdf-interest@w3.org>
On Fri, 23 Aug 2002, Paul Prescod wrote: > This conversation has gotten WAY off of my original statement. > > I said that: > > IFF RDF is not intended for human consumption, THEN it should not be > built upon XML. Because then the combined RDF/XML language will have > traded efficiency for *nothing*. Yup. I believe the RDF/XML syntax was designed in the expectation of human readers/writers, but that it was not optimised for such purposes. It is (was) a "good enough" format for shipping RDF around the network. Maybe I've been reading it for too long, but I don't find it too bad. If you take care to use typedNodes, and stick to the case conventions for classes and properties, it's not too hard to read once you get the 'striping' thing. I certainly much prefer reading an unknown chunk of XML data that uses RDF/XML conventions for its syntax to reading an unknown chunk of XML that uses its author's instincts about mapping abstractions into infosets. That said, I accept that RDF/XML might have been more intuitive, readable etc., had that been a priority of the original RDF Model and Syntax Working Group. Like all WGs, M+S had to juggle priorities... RDF/XML was designed with *both* people and machines in mind. It was designed and refined by a committee, and finalised before we had a great deal of implementor feedback and experience. Some members of the original WG cared a great deal about the XML-readability of the syntax; others preferred to prioritise work on other aspects of the spec (ie. the abstract data model), on the basis that we could always explore improved syntaxes at a later date. I believe the M+S WG found a reasonable balance between those two concerns, although there were of course bugs in both syntax and the graph abstraction. We could argue over whether RDF/XML was designed for machines, humans, or a combination; and we could engage in historical research or cry over spilt milk. The answer would be an elaboration of the obvious: it was a bit of both. We didn't have a very formal requirements process for RDF; we had a working group composed of individuals with a nice mix of agendas, skills and priorities. The RDF/XML syntax in the '99 REC emerged as a pragmatic outcome from the work of those individuals. Human-readability was one of several (somewhat implicit, often competing) goals for RDF, and RDF/XML. It wasn't the only one; however the tradeoffs and competition might be best understood in terms of WG logistics than in purely technical terms. There was a certain amount of time, effort, attention available; some but not all of that attention was spent refining the XML syntax. Now that we have more attention to spare, and a wider developer community, we can look at building on the work that was done for RDF 1.0... That seems more useful than worrying excessively about what the original intentions between RDF'99 were... Dan -- mailto:danbri@w3.org http://www.w3.org/People/DanBri/
Received on Friday, 23 August 2002 17:50:04 UTC