[Om3] [Om] Ontology for OpenMath CDs, SWiM as a CD editor

[Christoph LANGE]

Dear James, dear all,

  as there was no further reply about the OpenMath CD ontology (except
James'), I checked it into the OM3 repository at
https://svn.openmath.org/OpenMath3/owl.  Note:

* So far it's mostly OpenMath 2, except that it already supports notation
  definitions. I am, however, planning to extend it to OpenMath 3. Is there
  already any consensus about what elements will be added to the OM 3 CD
  language and what semantics they will have? And some draft document where I
  could read more about this?
* I found the directory name owl appropriate, in analogy to "xsl" and "sts".
  If you think it's not, feel free to move it elsewhere, but let me know
  beforehand, as SWiM depends on the ontology.
* A basic documentation is available as a README file.
* I have created a component "Ontology" in the OM3 Trac, where you can file
  bugs with the ontology.

On Saturday 14 June 2008 08:01:57 jhd at cs.bath.ac.uk wrote:
> On Fri, June 13, 2008 9:31 pm, Christoph LANGE wrote:
> > Have a look into
> > https://svn.salzburgresearch.at/svn/kiwi/IkeWiki/branches/SWiM/trunk/WEB-
> >INF/ontologies for the current state of the ontology I modeled for
> > OpenMath 2 CDs. The *.n3 file is human-friendly, the *.owl file generated
> > from that is
>
> Not, I'm afraid, for this human.

Sure, one needs to get used to it. But it's no rocket science. Semantic web
ontologies are commonly written in RDF. RDF is basically a language for
modelling (pseudo)graphs, where in our case CDs, symbol definitions, FMPs etc.
are types of nodes, and where edges can have types too. RDF is written down as
<subject,predicate,object> triples, and in most notations of RDF triples with
common subject or common <subject,predicate> are grouped together.

For more human-friendliness I recommend installing the ontology editors
Protégé or SWOOP and then opening the openmath.owl file with them. I could use
them for editing altogether, but then it would be hard to collaborate on the
source file via Subversion, as diffs would become hard to read.

> > The basic idea is to abstract a bit from the CD XML syntax and to make a
> > bit more of the semantics explicit than a RELAX NG schema can do. (E.g.
> > that the SignatureDictionary/@type attribute is not just a string or
> > URIRef or whatever, but in fact a link to a CD.)
>
> This seems like an excellent goal, though it is a much deeper level:
> semantics rather than syntax, I would say.

Thanks, James, that's it. I was a bit too modest…

Cheers,

Christoph

-- 
Christoph Lange, DERI Galway/Jacobs Univ. Bremen, http://kwarc.info/clange

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
Url : http://openmath.org/pipermail/om3/attachments/20080623/c4476c14/attachment.pgp