Re: My task from last week: Semantic free identifiers

It is a burden to carry this prefix file and maintain it, but its better
than not being able to do anything about it.

...from phone

On Jun 20, 2011 3:47 AM, "James Malone" <malone@ebi.ac.uk> wrote:

Hi,

On last week's call I was tasked with contacting Chris and Alan regarding
the form of the Relations Ontology URIs. There was some concern that the
current text fragment (e.g. part_of) would be replaced with a semantic free
identifier (e.g. RO_0000001). They confirmed this is the case. The
discussion stemmed from concern that our Atlas Triples use predicates which
are also in this form and there was a concern that it makes querying harder.
Alan Ruttenberg had the following useful suggestion (I would also be keen to
know if this resolves the concerns on the call or if I missed something more
fundamental)

for sparql they can write:

prefix part_of:<http://purl.**obolibrary.org/obo/RO_0000000<http://purl.obolibrary.org/obo/RO_0000000>
>

then query:

?a part_of: ?b

SPARQL processors should allow something like:

@include_prefixes<http://foo.**org/ro_prefixes.txt<http://foo.org/ro_prefixes.txt>
>

Cheers,

James



-- 
European Bioinformatics Institute,
Wellcome Trust Genome Campus,
Cambridge, CB10 1SD,
United Kingdom
Tel: + 44 (0) 1223 494 676
Fax: + 44 (0) 1223 494 468