Welcome Leo, Tenor update (was CFP: Semantic Desktop WS 2005)

[Leo Sauermann]

Hallo Tenor,


>I should take a moment to introduce Leo to the list and give a brief update on 
>some of the stuff that's been going on in the Tenor world (since I've been 
>mostly quiet here).
>  
>
Thanks, more about me can be found on
http://www.dfki.de/~sauermann


>In the stuff that I'm planning on checking in the basic components -- Nodes, 
>Links and Property creation is there and mostly works.  The components for 
>searching and iterating over the graph aren't in place though those will 
>follow.  
>
I still want to point out the similarities to RDF and the corresponding 
frameworks.
In RDF (Semantic Web technology by W3C) the core are "resources" 
(=Nodes) and relations between resources,
named "statements" (=Links)

RDF uses namespaces to identify different link types and uses URIs to 
identify the nodes.
A possible notation for RDF for serialisation or storage is RDF/XML or N3.
http://www.w3.org/DesignIssues/Notation3.html
http://www.w3.org/2000/10/swap/Primer

for example, if one file is related to another file, you could note the 
link like this:

  <file:///home/leo/docs/filea> 
<http://purl.org/dc/elements/1.1/relation> 
<file:///home/leo/otherdocs/otherfile>.

one can abbreviate the namespace used for the relation like this:
@prefix dc:
      <http://purl.org/dc/elements/1.1/>.
<file:///home/leo/docs/filea> dc:relation 
<file:///home/leo/otherdocs/otherfile>.

There exist vast amount of RDF test data on the web and schemas (= 
metadata vocabularies) on www.schemaweb.info
That would allow compability to other solutions, not only desktop but 
email clients, social networking software, websites, etc.

>Specifically one of the next components to be written is a graph 
>iterator of some sort that is aware of its position in graph traversal and 
>avoids the requirement to do complete traversal of a query in one step or 
>hard-limiting traversals.  Also each of the directories (currently tenor -- 
>the library, storage -- db stuff, test, docs and apps) also have README files 
>outlining the components that need to be filled in.
>  
>
You don't have to do this.
If you decide to use RDF, you would have stable and tested code 
available that does all that, which would spare some sweat on ugly stuff.
An implementation that can do the required things is

http://librdf.org/ - Apache or LGPG license, written in C, mappings to 
all major programming languages (java, python, c#, objC ....)

http://rdfstore.sourceforge.net/ - Apache or BSD license, RDF database 
with Query and traversal support. written in, C, needs a database

http://sourceforge.net/projects/threestore/ - GPL, C written. Database

there are tons of projects in other programming languages out there, see
http://www.ilrt.bris.ac.uk/discovery/rdf/resources/#sec-tools


-> If someone wants to know more about how we use this RDF stuff to link 
and index desktop data, please contact me, I have been hacking this 
stuff for years now.


cheers
Leo