Re: Re: HCLSig July 6 Teleconference Agenda

>
> >1. You may use owl:sameAs to indicate the same experiment published
> >in two different ontologies.
>
> I think you are missing my point. We have other ontologies, like the
> biological pathways exchange language (BioPAX), that describe biological
> information -- in the case of BioPAX it is molecular interactions and
> pathways. This ontology does NOT describe an experiment. Datafiles that use
> this ontology contain instances of such classes like "protein",
> "molecularInteraction", "pathway". These are not experiments.
> You should find a way to connect (with an edge between two nodes) such
> RDF/OWL graphs to the RDF/OWL of your experiment ontology. Your ontology
> should, in my opinion, define a RDF property that can act as such a
> connection.


Thanks for correcting me. Your suggestion to have a property to  uniquely
identify (and correlate) the resource of main concepts of an experiment.
This is similar to correlate the materials and tools used in the protocols
or procedures.  The "Protocol" class has an property called "productUsed",
which should have value as URI of the product (reagent or any tool).  If
product providers can define and publish their product resources, then
anyone will know exactly what tools were used by the experiment assuming the
tools are listed for the published experiment.

>2. SPE uses many properties defined in FOAF, but I find it
> >difficult to re-use its classes like Person, Group, Organization
> >because they miss some of the properties required by SPE's
> >corresponding classes.How do you re-use a FOAF class if the class
> >you need has additional properties?
>
> You could simply add the properties you need without a problem. Or (if you
> want to leave the original FOAF class untouched for some reason), you could
> define a subclass for each of the classes of FOAF, and make your additions
> to these subclasses. (e.g. a subclass "scientific-project" of
> "foaf:project", a subclass "researcher" of "foaf:person" and so on).


This is what I thought. In any case, a new class is needed to make clear
what extra properties it has.  Since the scope of this task is to provide
general terms for all research areas, scientifc or social or political or
whatever, better choice of class names are "Project", "Researcher",
"Experiment", etc.

AJ

Received on Friday, 7 July 2006 20:10:15 UTC