RE: API for RDF: locutor

Here here! In particular the notion of ordering in a non-order-dependent
construct bugs me. Imagine I create some RDF with foo as _1 in some bag, bar
as _2, and so forth, then put it into a database, then take it out of the
database back into RDF, but the database serializes it with bar coming
before foo. 


-----Original Message-----
From: Jan Grant []
Sent: Thursday, March 09, 2000 1:40 AM
Subject: RE: API for RDF: locutor

On Wed, 8 Mar 2000, Jan Grant wrote:

> 3. It's possible to envisage a higher-level API that builds on a
> fundamental RDF API to support easy access to locutors- ie,
> several levels of API that add value to underlying levels. Bear in mind
> though that these are only an interface: the RDF model they present does
> not have to be implemented (nor should it be) naively.
>  However, it would be up to the implementation to 'fake' the existence
> of RDF triples that are implicit in its locutor-friendly
> implementation; since people can always drop back to the basic "ask
> about the triples" style of querying the RDF model.

However, I would hasten to add that support for bag, sequence,
alternative (and sets, sigh) membership should definitely be added to
the API. The representation of these ideas in triples seems clunky at
best and probably only useful for serialising RDF; I'd hate to think
that people really were testing for A -_1-> B to determine bag
membership :-(

jan grant, ILRT, University of Bristol.
Tel +44(0)117 9287163 Fax +44 (0)117 9287112 RFC822
Whenever I see a dog salivate I get an insatiable urge to ring a bell.

Received on Thursday, 9 March 2000 14:39:28 UTC