W3C home > Mailing lists > Public > www-rdf-comments@w3.org > October to December 1998

SiRPAC bug report: relative URIs unresolved

From: Dan Brickley <Daniel.Brickley@Bristol.ac.uk>
Date: Mon, 21 Dec 1998 20:20:45 +0000 (GMT)
To: www-rdf-comments@w3.org
Message-ID: <Pine.GHP.4.02A.9812212011040.14803-100000@mail.ilrt.bris.ac.uk>


The RDF M&S spec defines triples as triple(res,property,value) where
value may be a URI reference or a literal, and res and property are URI

The output from SiRPAC doesn't always produce URIs in the appropriate
places: for example, generated node identifiers (genid_6 etc) for
"anonymous" nodes. Also, relative URIs (and namespace refs, though I'm
unsure about legality of trying that!) and resources identified only by
ID within the RDF/XML do not produce triples that contain URIs.

I'm not sure what's best to do here. Seems a bit over-the-top to create
uuid: URIs for each anonymous node. The URI specification[1] defines
four rules for derriving absolute from relative URIs + context, and
includes warnings about 'meaningless' relative URIs. It includes
provision for 'application specific' rules for creating full URIs as a
last resort. This might be worth considering in cases where the RDF/XML
doc being parsed has no obvious URI.

Perhaps an additional parameter for SiRPAC might be 'baseURI', so that
the calling context could pass in a URI against which relative URIs
should be considered?



[1] http://www.ics.uci.edu/pub/ietf/uri/rfc2396.txt

Institute for Learning and Research Technology http://www.ilrt.bris.ac.uk/
University of Bristol,  Bristol BS8 1TN, UK.   phone:+44(0)117-9288478
Received on Monday, 21 December 1998 15:20:49 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:15:12 UTC