Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[bioontology-support] Three things

Alan Ruttenberg alanruttenberg at gmail.com
Thu Aug 11 09:46:41 PDT 2011


On Wed, Aug 10, 2011 at 8:28 PM, Ray Fergerson
<ray.fergerson at stanford.edu>wrote:

> Alan,
>
> We believe that we have fixed the processing of the metrics for OBO
> ontologies and that the numbers visible in production for "classes with no
> definition" are now correct. Please let us know if this turns out to not be
> the case.
>

It looks better, but still suffers from the issue that OBI imports a number
of terms from other ontologies and the metrics should really be split out
into what OBI does versus information about the imported stuff.

All the imported terms are either explicitly in an imported file, or single
imported and marked with an annotation property 'imported from'
http://purl.obolibrary.org/obo/IAO_0000412

In addition it still reports incorrect information when the ontology editor
doesn't fill in the special properties. This shouldn't be a problem when a
subsequent version is uploaded, but means that people who haven't done that
- who uploaded a file before metric were in place - still get 0 authors. I
think that's wrong. Metrics should only be computed when you know they will
be correct. For the moment, I think it would be prudent to have metrics
computation be an opt-in. Finally, when updating the metadata, if the values
of the annotation properties change, the metrics should be hidden, or marked
as out of date, until such time as they are recomputed. For example, I
updated the IDO metadata to include the IAO properties but it still shows
that all classes do not have a name or author.

As far as OBO goes, it would be nice if you augmented your defaults to both
look for the current defaults as well as the OBO defaults.

Label:
rdfs:label ok

Synonym:
http://www.w3.org/2004/02/skos/core#altLabel OR
http://purl.obolibrary.org/obo/IAO_0000118

Note that alternative labels are all direct values of IAO_0000118 as well as
any that are marked as annotation subproperties of it, transitively. Also
note that it is not uncommon to have more than one property to record
synonyms and you might want to provide this facility to your users.

Definition:
http://www.w3.org/2004/02/skos/core#definition OR
http://purl.obolibrary.org/obo/IAO_0000115

Author:
http://purl.org/dc/elements/1.1/creator OR
http://purl.obolibrary.org/obo/IAO_0000117

The OBO ontologies, which will all eventually adopt these properties, is a
large enough community that I
think it warrants such a change.

Incidentally, use of dc:creator as author property is incorrect. If you read
the documentation
you will see that the thing created is what the term means, not the term
itself. So for a term that
represents a sculpture, the dc:creator would be the person who make the
sculpture, not the person
who supplied documentation about it for the representation.

BTW, another property that is sorely missed is license information. People
need to know under what conditions
they can use the ontologies and this is difficult to find at the moment. A
URL value for the license would be adequate.

As far as metrics go, putting in separate statistics for the imported terms
should be considered as we want to
encourage reuse. To that extent all the statistics you compute for the
ontology would be appropriate to compute for the
imported terms, as well as some others, such as how many of the imported
terms are used in an axiom in the importing
ontology.

-Alan



>
> Ray
>
> > -----Original Message-----
> > From: bioontology-support-bounces at lists.stanford.edu
> > [mailto:bioontology-support-bounces at lists.stanford.edu] On Behalf Of
> > Alan Ruttenberg
> > Sent: Sunday, July 17, 2011 7:35 AM
> > To: Trish Whetzel
> > Cc: support at bioontology.org Support
> > Subject: Re: [bioontology-support] Three things
> >
> > Doesn't appear to me - I don't see the administrative tab.
> >
> > Do you plan to respond to the other issues I raised?
> >
> > -Alan
> >
> >
> > On Sun, Jul 17, 2011 at 9:34 AM, Trish Whetzel <whetzel at stanford.edu>
> > wrote:
> > > Your account "alanruttenberg" is set to manage IAO. The BioPortal
> > PURL is currently set-up to resolve to the visualization of the latest
> > version of the ontology.
> > >
> > > Trish
> > >
> > >
> > > On Jul 11, 2011, at 6:58 PM, Alan Ruttenberg wrote:
> > >
> > >> 1) Can you please refrain from showing metrics on existing
> > ontologies
> > >> that have not updated their metadata, or to include a disclaimer
> > >> warning that the metrics may not be valid. I have looked now at
> > IAO's
> > >> page and find once again that the metrics are bogus. I have noted
> > >> this before and am surprised that no action was taken considering
> > the
> > >> importance the Bioportal team must give to presenting accurate
> > >> metrics, not to mention the potential damage to the reputations of
> > >> the ontology developers who the Bioportal aims to support.
> > >>
> > >> 2) Can you include some information on the details page identifying
> > >> the party able to edit the summary? I don't know what account it is
> > >> under - I have tried two likely candidates now and neither works -
> > my
> > >> own "alanruttenberg" and the obi account
> > >> "ontology.bio.investigations". In the case of IAO can you please let
> > >> me know which account has management ability?
> > >>
> > >> 3) Can you clarify what the bioportal PURL is? Apparently it leads
> > to
> > >> the ability to explore the current version. In that case it should
> > be
> > >> labeled as such: "Link to visualize" or "Bioportal visualization
> > purl"
> > >> or somesuch. In addition, the ontology IRI and version IRI (you
> > know,
> > >> the IRI's that the OWL spec considers normative  -
> > >> http://www.w3.org/TR/2009/REC-owl2-syntax-
> > 20091027/#Ontology_IRI_and_
> > >> Version_IRI) should be given more prominence so as not to mislead.
> > >>
> > >> Thanks,
> > >> Alan
> > >> _______________________________________________
> > >> bioontology-support mailing list
> > >> bioontology-support at lists.stanford.edu
> > >> https://mailman.stanford.edu/mailman/listinfo/bioontology-support
> > >
> > >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/bioontology-support/attachments/20110811/0b665c03/attachment.html>


More information about the bioontology-support mailing list