Re: Manchester Syntax and labels

I think you said during the meeting that OBO was planning to make sure
their labels were unique/unambiguous and stable.  In that case, I'd
suggest they go ahead and make text that is currently in labels be part
of the IRIs, in the OWL versions of the ontologies.

This doesn't solve the general problem -- using ontologies in multiple
languages, for instance -- but maybe it doesn help bridge the gap
between OBO and OWL.    

The tool for doing the conversion/migration would be more-or-less the
tool we talked about as the "post-processing" approach -- something
which converts between an RDF graph using ugly URIs+labels and an RDF
graph using nicer URIs.

     -- Sandro

> For discussion of the issue of using labels in Manchester syntax, here
> is an example from the OBI ontology, as rendered in protege 4, and is
> typical of what you can expect in OBO ontologies. The OBO ontologies
> are the largest collection of Biomedical ontologies there is.
> 
> With labels
> 
> Assay:
> achieves_planned_objective some 'assay objective'
> has_specified_input some (material_entity
>                           and has_role some 'evaluant role')
> has_specified_output_information some ('information content entity'
>                                        and 'is about' some (continuant
>                                                              and
> has_role some 'evaluant role'))
> realizes some 'evaluant role'
> has_specified_output only Nothing
> 
> Without labels
> 
> OBI_0000070:
> OBI_0000417 some OBI_0000441
> OBI_0000293 some (MaterialEntity
>                   and OBI_0000316 some OBI_0000067)
> OBI_0000301 some (IAO_0000030
>                   and IAO_0000136 some (Continuant
>                                          and OBI_0000316 some OBI_0000067))
> OBI_0000308 some OBI_0000067
> OBI_0000299 only Nothing
> 
> -Alan

Received on Thursday, 2 April 2009 01:07:24 UTC