W3C home > Mailing lists > Public > public-rdf-wg@w3.org > February 2013

RE: Problem with auto-generated fragment IDs for graph names

From: Markus Lanthaler <markus.lanthaler@gmx.net>
Date: Wed, 20 Feb 2013 15:03:17 +0100
To: "'Andy Seaborne'" <andy.seaborne@epimorphics.com>
Cc: <public-rdf-wg@w3.org>
Message-ID: <022401ce0f73$09d14b50$1d73e1f0$@lanthaler@gmx.net>
On Wednesday, February 20, 2013 2:32 PM, Andy Seaborne wrote.

> > You can easily rename each bnode identifier. Even the ones that are
> already
> > there. You can't do the same with IRI (fragments).
> 1/ In this case, with the base decided by the parser, why not?  This is
> only document scope.

Because that's not always true. Bnodes would solve the problem in all
instances, base-less framentIds just in Manu's very specific use case. How
would you do it if the document is not transient but is hosted somewhere
(and thus has a base)?

> 2/ They are generated, not used, so the parser can generate unique
> strings.
> For unique bnode ids or fragments, you have to look at the whole doc
> first to know the string is unique.

Not necessarily.. you can also relabel them in a streaming-manner. You just
need to remember the mappings (old->new).

I think the crucial point here is that no global identifiers (IRIs) should
be minted automatically. Bnodes are document-local by definition. IRIs are
global by definition. Fragment identifiers without base are strings, not
IRIs, and thus not valid in (abstract) RDF.

Markus Lanthaler
Received on Wednesday, 20 February 2013 14:03:47 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:02:10 UTC