- From: Tim Berners-Lee <timbl@w3.org>
- Date: Fri, 28 Feb 2003 11:42:53 -0500
- To: <www-archive@w3.org>
- Cc: "Dan Connolly" <connolly@w3.org>, "Eric Miller" <em@w3.org>
It has been suggested that I give my take on the meaning of an RDF document. THEN NEED FOR SPECIFYING THE MEANING OF DOCUMENTS The web is built as a communications medium by virtue of a stack of common specifications. http://www.w3.org/DesignIssues/Stack So, while the content of the web is very decentralized and grows in an uncontrolled way, this whole decentralized system depends on a rather mimuimal but essential set of globally agreed standards. To use the Web is to agree implicitly to follow these standards. That's how it works. When a language specification is added to the specs, the goal of the specification is to provide a step in the stack. Any recipient/reader of a message in the language, should, when equipped with the specification, be able to determine the meaning of the message. I use these terms loosely, and most of the layers of the Web specification stack have to date been defined by engineers with a good deal of thought. There have been formal models in relatively rare cases, though syntaxes are defined by grammars, and protocols define message sequnces with finite state machines and so on. The OWL terms, beccuase they are purely mathematical, can be defined in terms of axoims or model theory. Some langauges are meta-languages in that they create a framework with which to define further languages. XML and RDF are examples. In this cases, the metalangauge spec hands off to a langauge spec in a well-defined way. XML hands off through the namespace identifier. RDF has to define how to figure out what an RDF document means. THE MEANING OF AN RDF DOCUMENT The meaning of an RDF document I think is well defined in practice, in that it is the combined (conjunction) of the meanings of each statement, and the the meaning of a statement is defined by the definitition of the predicate. The subject, predicate and object are all IDENTIFIED by a URI, but the meaning of the statement if defined by the authority for the predicate, with the subject and object being parameters to that definition. That may be for some quick definition without much explanation, The RDF core group does *NOT* have to concern itself with how URIs for precicates are allocated,or how specs are defined. It does not have to model the social contexts in which RDF documents have found themselves. I think some people have been concerned that the specification in the RDF spec of the meaning of RDF documents would somehow change or redefine the social contexts in which email messages are sent and web pages are written. This is not so. That is outside the scope of a the group. timbl
Received on Friday, 28 February 2003 12:14:14 UTC