Update of EvoOnt's diagram needed ?
Olivier Berger
I've had a look at the ontology's docs and the diagram on the front page
(http://www.ifi.uzh.ch/ddis/typo3temp/pics/939c121d4a.png) ... and I
think some update may be needed...
Maybe just having a number in the ontology would help identify if it's
up to date or just an old version not authoritative, but for
information.
Hope this helps,
Best regards.
--
Olivier BERGER <olivier...@it-sudparis.eu>
http://www-public.it-sudparis.eu/~berger_o/ - OpenPGP-Id: 1024D/6B829EEC
Ingénieur Recherche - Dept INF
Institut TELECOM, SudParis (http://www.it-sudparis.eu/), Evry (France)
Henry Story
On 27 Jan 2009, at 18:10, Olivier Berger wrote:
> Hi.
>
> I've had a look at the ontology's docs and the diagram on the front
> page
> (http://www.ifi.uzh.ch/ddis/typo3temp/pics/939c121d4a.png) ... and I
> think some update may be needed...
baetle wiki, and subversion repository to add some thing.
Thanks for the pointer.
I had not realised there were so many relations in BOM.
It seems to me that it is unecessary to create an inverse for every
relation. In N3 one can write
x is rel of y .
which means
y rel x .
So having two names places a lot more burden on inference engines,
which at present are just not all that good. This criticism would also
apply to sioc, as they have done this a lot.
> Maybe just having a number in the ontology would help identify if it's
> up to date or just an old version not authoritative, but for
> information.
the latest documentation on them.
Henry
Olivier Berger
>
> On 27 Jan 2009, at 18:10, Olivier Berger wrote:
> > Hi.
> >
> > I've had a look at the ontology's docs and the diagram on the front
> > page
> > (http://www.ifi.uzh.ch/ddis/typo3temp/pics/939c121d4a.png) ... and I
> > think some update may be needed...
>
> Olivier, if you have a gmail account, I can give you access to the
> baetle wiki, and subversion repository to add some thing.
>
Sure, why not (email sent off-list).
> Thanks for the pointer.
> > Maybe just having a number in the ontology would help identify if it's
> > up to date or just an old version not authoritative, but for
> > information.
>
> Yes, we need pointers on the baetle wiki to the latest ontology, and
> the latest documentation on them.
>
Actually, it was mainly addressed at people in Zurich for contents found
at http://www.ifi.uzh.ch/ddis/evoont.html ... but improving the wiki on
googlecode site wouldn't hurt, I suppose.
Regards,
Olivier Berger
>
> > Hi.
> >
> > I've had a look at the ontology's docs and the diagram on the front
> > page
> > (http://www.ifi.uzh.ch/ddis/typo3temp/pics/939c121d4a.png) ... and I
> > think some update may be needed...
>
> Olivier, if you have a gmail account, I can give you access to the
> baetle wiki, and subversion repository to add some thing.
Oh, also, it seems the versions of EvoOnt ontologies in the svn at
http://code.google.com/p/baetle/source/browse/#svn/evoont/trunk are
outdated too... is this bug/feature ? ;)
I would understand that UZH people prefer to have the reference version
on their side... but then this should probably be explicitely mentioned
as a copy on code.g.c, then...
My 2 more cents.
Regards,
Henry Story
On 27 Jan 2009, at 18:46, Olivier Berger wrote:
> Le mardi 27 janvier 2009 à 18:28 +0100, Henry Story a écrit :
>>
>> On 27 Jan 2009, at 18:10, Olivier Berger wrote:
>>> Hi.
>>>
>>> I've had a look at the ontology's docs and the diagram on the front
>>> page
>>> (http://www.ifi.uzh.ch/ddis/typo3temp/pics/939c121d4a.png) ... and I
>>> think some update may be needed...
>>
>> Olivier, if you have a gmail account, I can give you access to the
>> baetle wiki, and subversion repository to add some thing.
>
> Oh, also, it seems the versions of EvoOnt ontologies in the svn at
> http://code.google.com/p/baetle/source/browse/#svn/evoont/trunk are
> outdated too... is this bug/feature ? ;)
openly as a group.
But that did not materialize.
> I would understand that UZH people prefer to have the reference
> version
> on their side... but then this should probably be explicitely
> mentioned
> as a copy on code.g.c, then...
relevant now being out of date and not the
core element.
For EvoOnt one should point to their core namespace.
Henry
Abraham Bernstein
Jonas, who is in charge of EvoOnt is off-line until the second week of
February. I am sure he will take care of things when he is back. I
vaguely remember he had some rationale for the many relationships. But i
do not remember what they were.
Best
Avi (Bernstein)
-----------------------------------------------------------------
| Professor Abraham Bernstein, PhD
| University of Zürich, Department of Informatics
| web: http://www.ifi.uzh.ch/ddis/bernstein.html
Olivier Berger
I have worked on an SVG export of the original file provided by Jonas
Tappolet (originally in Visio), and have updated it to (hopefully) fix
some small issues that were present in the previous diagram.
I have only reviewed bom.owl so there may be some problems still, in the
SOM and VOM models. Also the BOM part of the diagram is a bit
overcrowded now, but I preferred not to change it too much.
I've committed the SVG (which was edited with Inkscape, so I hope it is
correct on any platform) into
http://code.google.com/p/baetle/source/browse/#svn/evoont/trunk
Here's the corresponding PNG version :
http://baetle.googlecode.com/svn/evoont/trunk/evoont.png
There are some small issues in BOM, I'm afraid, but I'll deal with that
in another email.
Hope this helps, and you don't mind the commit directly in the baetle
SVN, in the hope some others could improve it.
Best regards,
Henry Story
On 12 Mar 2009, at 16:42, Olivier Berger wrote:
> Here's the corresponding PNG version :
> http://baetle.googlecode.com/svn/evoont/trunk/evoont.png
Very helpful Olivier. I just pointed some group working on some
related topic to this.
By the way, the mime type is incorrect:
$ curl -I http://baetle.googlecode.com/svn/evoont/trunk/evoont.png
HTTP/1.1 200 OK
Date: Thu, 26 Mar 2009 10:56:00 GMT
Server: Apache
Last-Modified: Thu, 12 Mar 2009 15:35:45 GMT
ETag: "206//evoont/trunk/evoont.png"
Accept-Ranges: bytes
Content-Length: 718929
Content-Type: application/octet-stream
I think you can set the mime type in svn. I did that for the
ontologies we published there. It's a piece of metadata.
I think your diagram would be a lot clearer if you removed all the
inverse relations from it. No need for those really. I think it is an
error in sioc to have put so many in there. Every relation has an
inverse, so there is no need to define them really.
N3 has a nice syntactic sugar for those:
<http://bblfish.net/> if foaf:homepage of :bblfish .
Henry
Olivier Berger
>
> By the way, the mime type is incorrect:
>
> $ curl -I http://baetle.googlecode.com/svn/evoont/trunk/evoont.png
> HTTP/1.1 200 OK
> Date: Thu, 26 Mar 2009 10:56:00 GMT
> Server: Apache
> Last-Modified: Thu, 12 Mar 2009 15:35:45 GMT
> ETag: "206//evoont/trunk/evoont.png"
> Accept-Ranges: bytes
> Content-Length: 718929
> Content-Type: application/octet-stream
>
> I think you can set the mime type in svn. I did that for the
> ontologies we published there. It's a piece of metadata.
>
I did :
svn propset svn:mime-type image/png evoont.png; svn commit evoont.png
and that did the trick.
Thanks for noticing.
> I think your diagram would be a lot clearer if you removed all the
> inverse relations from it.
Maybe.
> No need for those really. I think it is an
> error in sioc to have put so many in there. Every relation has an
> inverse, so there is no need to define them really.
I think it may not be that useful to have them on diagrams, but for the
contents of the ontology, I think they are probably necessary so that
the RDF documents can be read/written by humans with the most
significant "direct" phrasing instead of using "indirect" descriptions.
For instance if I describe a parent/children relashionship, and I say
<Rene> name "Rene"
<Rene> hasChild <Olivier>
it may be much more convenient than having to say :
<Rene> name "Rene"
<Olivier> hasParent <Rene>
Sorry if this is obvious (or obviously wrong)... my SW knowledge is
improving every day ;)
>
> N3 has a nice syntactic sugar for those:
>
> <http://bblfish.net/> if foaf:homepage of :bblfish .
Uh... I must say I don't get that one... care to elaborate ?
Thanks in advance.
Regards,
Henry Story
On 26 Mar 2009, at 13:51, Olivier Berger wrote:
>> N3 has a nice syntactic sugar for those:
>>
>> <http://bblfish.net/> if foaf:homepage of :bblfish .
Sorry my bad
<http://bblfish.net/> is foaf:homepage of <http://bblfish.net/people/henry/card#me
> .
is just the same as
<http://bblfish.net/people/henry/card#me> foaf:homepage <http://bblfish.net/
> .
See:
http://www.w3.org/2000/10/swap/doc/Overview.html
You can then use cwm if you want to transform it into rdf/xml.
To write code, use N3, to publish it rdf/xml - at present, until a
better standard gets a blessing.
Henry