- From: Christian de Sainte Marie <csma@ilog.fr>
- Date: Tue, 24 Jul 2007 16:05:44 +0200
- To: Dave Reynolds <der@hplb.hpl.hp.com>
- CC: Gary Hallmark <gary.hallmark@oracle.com>, RIF WG <public-rif-wg@w3.org>
Dave Reynolds wrote: > Gary Hallmark wrote: >> >> Xpath expressions returns a list of nodes, and nodes can be elements, >> attributes, and typed literal values. So we'd need at minimum some >> kind of list in RIF (which we've talked about quite a bit), and I >> don't know what you do about elements and attributes. Forbid them? >> Introduce an opaque nodeId? I had the operational semantics of production rules in mind when I wrote that, where only ground rules (rule instances) are evaluated (that is, with all variables bound before). And I had XPath (or a subset of XPath) as a syntax to navigate the data when the data model (that the rule interchange refers to) is XML. Not necessarily for application to an XML data source. > So how do existing rule languages handle this at the moment? > Does everyone have their own data modelling language which they map XML > onto? If so what do those modelling languages look like? The processing model I had in mind is that both end of the interchange have their own data modeling language (maybe the same, but that is irrelevant : the point is that this should be transparent for the interchange). And if they both know the data model that the interchanegd RIF document refers to (which is a prerequisite for the interchange), it means that they have a mapping between that data model and their own. And, thus, they will know how to map a in the data that is specified according to the common data model into a path specified according to their own data model. Christian
Received on Tuesday, 24 July 2007 14:09:47 UTC