Re: Biological and other non-model citations in CellML metadata?

On Fri, 30 Mar 2007, Matt  wrote:

>>> In those cases, for example bqs, where the object of the reference is
>>> indeterminate or not of interest, it would be helpful to be able to
>>> filter metadata without having to have knowledge of the values (in
>>> this case reference URI schemes or identifiers) to determine the type
>>> of metadata property you are dealing with.
>>
>> How do-you do that? How do-you decrypt a foreign language without a
>> dictionary?
>
> You have typed properties, for example you simply pass over everything
> that is a bqs:reference because you are not interested in it. You
> don't need to know anything about the schema restrictions underlying
> objects of bqs:reference to dertermine that this is a bqs:reference.
>
> With isDescribedBy being used generally for all kinds of external
> references, you have no way of knowing what the intention of the
> instance of this relation without inspecting the object value, and to
> do that you need to understand all current and future kinds of data
> that could be used.

I think this is where the misundersting is. I thought you had a
problem with the use of URIs, while you have a problem with the
qualifier.

> What do you mean by hardcode it in your language?

If you create an element <bqs:PubMed_id> in your language, rather than
having a generic reference scheme, with a type PubMed defined elsewhere.

>> The big advantage of externalising the type of metadata is that the
>> scheme is generic.
>
> What do you mean by the 'type' of metadata?

EC page, PubMed entry, DOI indexed document, UniProt entry, Gene
Ontology term etc.

> I think more I am misunderstanding the range of use isDescribedBy is ok for.

isDescribedBy is a relationships. It does not tell anything about the
component to be annotated, or the resource used to annotate it. It
describe the relationships between the concept represented by the
component and the concept or object represented by the resource. It is
similar to the Dublic-core isVersionOf, hasVersion etc.

In the initial annotation scheme, we used only dc:relation. Then we
decided to enrich the relationships by using the dcterms. But 1) they
were too limited, 2) their semantics was slightly different
(dcterm:hasPart really meant that the annotated document was
physically composed of the parts pointed to).

-- 
Nicolas LE NOVERE,  Computational Neurobiology,
EMBL-EBI, Wellcome-Trust Genome Campus, Hinxton, Cambridge, CB10 1SD, UK
Tel: +44(0)1223494521, Fax: +44(0)1223494468, Mob: +44(0)7833147074
http://www.ebi.ac.uk/~lenov, AIM: nlenovere, MSN: 
nlenovere-PkbjNfxxIARBDgjK7y7TUQ@xxxxxxxxxxxxxxxx