[Xesam] Nepomuk/Desktop/Shared ontologies - svn layout as discussed at the GCDS last week

Sebastian Trüg strueg at mandriva.com
Mon Jul 13 07:10:31 PDT 2009

Hi guys,

as discussed last week at the GCDS I changed the layout of the ontology svn 
for the OSCAF project:

- trunk
  |- ontologies
     |- base
        |- dces.rdf
        |- dcq.rdf
        |- dctype.rdf
        |- rdf.rdfs
        |- rdfs.rdfs
     |- nrl
        |- nrl.trig
        |- doc
           |- nrl-header.html
     |- nao
        |- nao.trig
        |- doc
           |- nao-header.html
     |- ...
  |- draft
  |- tools

The 'ontologies' folder contains all "stable" ontologies which will make up 
our releases. The 'draft' folder contains ontologies under discussion. The 
'tools' folder will contain the scripts used to generate the docs and to 
verify commits (Ivan took responsibility for creating the scripts).

Each ontology has its own folder which contains the one trig (turtle + named 
graphs) file and the doc folder. The doc folder contains an html header which 
will be used to generate the ontology documentation in combination with the 
comments from the ontology itself.

The trig file contains two named graphs:
1. The ontology itself
2. the metadata graph which contains information such as the graphs' type, the 
   version, the state of the ontology, the last modification date, and so on.
   This information is encoded using NRL and NAO.

The 'ontologies/base' folder contains base ontologies which are not under our 
influcene, i.e. which we do not maintain but depend on. At the moment these 
are all serialized in rdf+xml since that is what everybody seems to use at the 

I moved the old content to branches/legacy. This also contains the java apps 
used to generate the docs before. Ivan, maybe those are of interest to you 
when creating the scripts.

Open issues are:
* We need the doc generation scripts and a build system (we decided to use 
* The base ontologies are encoded in rdf-xml instead of trig. Should we
  convert them to trig (and add metadata in the process and also have a 
  slightly more complex update mechanism if they change) or should we keep 
  them in their original form? The latter sparks the problem of serialization 
  again: if we only use trig client that need to parse the ontologies can 
  depend on that. If not, we need to somehow store which serialization is 
  used. In KDE I use desktop files for that. But there are probably other 
  possibilities, too.
* Copyright: none of the files contain a copyright header atm. We need to 
  settle on the licence and then add a header to all files. Also atm the 
  documentation stated that the ontos are (C) Nepomuk consortium which does 
  not exist anymore. Question is: who is it now?
  (I know we discussed this before but I fail to find the email at the moment. 
  I only Nepomuk integration in KMail was already done...)


More information about the Xesam mailing list