Re: [dotNetRDF-develop] Suggestion / Question about Euler-type reasoners and integration to other open projects

[Rob V. <rv...@do...>]

Hi Andres

I honestly don't know how easy it would be to do any of the integration and in some ways I'd rather do new implementations of things from scratch rather than trying to merge and shoehorn external code into our library.
As you'll see from our ideas page [1] we have various reasoning projects planned but no time to do them for the time being.

With regards to Euler I'm not sure how easy it would be to implement for dotNetRDF as I have never looked in detail at how the reasoning algorithm works.

As of our latest release (out today) we do support simple N3 rules of the following form:

{ ?x a ex:Car } => { ?x a ex:Vehicle }

To use these you can use the SimpleN3RulesReasoner [2] found under VDS.RDF.Query.Inference

If you really need/want Euler reasoning that I suggest talking to Jos de Roo about how his algorithm works and consider trying to implement it yourself using the dotNetRDF API and submitting it as a contribution to the project.

I see no value in merging SemWeb with dotNetRDF as there are few features it provides which we don't already have and/or do better.  If you disagree please let me know which feature from it you'd like to see in dotNetRDF and we can discuss whether that is feasible.

As for OWL reasoning none of these libraries (including dotNetRDF) support any form of OWL reasoning currently.  Like I say we plan on adding it in the future but can't give you any indication of timescale for this.

If your OWL reasoning needs are covered by Pellet and you can have all the data externally in a Pellet Server instance then you can use our built in Pellet Server client [3] to access some features of Pellet from within dotNetRDF.

Regards,

Rob Vesse

[1] http://www.dotnetrdf.org/content.asp?pageID=Ideas
[2] http://www.dotnetrdf.org/api/index.asp?Topic=VDS.RDF.Query.Inference.SimpleN3RulesReasoner
[3] http://www.dotnetrdf.org/api/index.asp?Topic=VDS.RDF.Query.Inference.Pellet.PelletServer

----------------------------------------
From: "Andres Hohendahl" <and...@gm...>
Sent: 01 March 2011 21:58
To: rv...@do...
Subject: Suggestion / Question about Euler-type reasoners and integration to other open projects 






Hello 
  
I have been 
searching on C# W3C-RDF/N3 libraries, and found yours, which appears 
to be very good, full, well fomatted and somehow extensible. 
I also found many 
truncated words like this 
http://eulersharp.sourceforge.net/
which claims to have 
a vert fast reasoner, but only a poor parser/object 
interfase. 
I tested it but the 
information is bad-formatted and difficult to understand, compiled and made a 
lot aof demonstrations, many of which dont work as claimed!. 
...but the reasoner 
works fine at least with simple deductions and samples. 
  
also found this 
one: 
SemWeb: A Semantic 
Web Library for 
C#/.NET
========================================== 
 
By Joshua Tauberer 
<http://razor.occams.info> 
 
http://razor.occams.info/code/semweb 
The code seems to be 
abandoned!  
  
and it seems to be 
fine, also well formatted and full of good-samples, but lacks a 
reasoner 
(only blame: its 
slow, because it don''t use .NET2.0 faster classes, and still hangs with 
ArrayList() instances and some nasty enumeratos, because the code is mostly 
ported from Java2 with an automatic converter)  
Also here: the code 
seems to be abandoned! 
  
  
So my questions 
are: 
  
I am developing an 
AI-dialog description language, with a runtime and a real compiler, for active 
and intelligent agents creation. 
I use reasoners in 
the old-way (neural nets, learners, bayesian inference engines, 
etc.) 
I found those 
AI-stuff work well but there is a lot more of knowledge out there, hidden in a 
bunch of standards, spread in differents databases and 
websites. 
I want to implement 
some W3C standard interoperative semantic-oriented tasks, mainly ontology 
reasoning for a NLP Query-Answering system 
I want to integrate 
wordnet and own ontologies also as an active ontology for 
reaoning. 
I want them to 
interact and reason with OWL FOAF KIT and other ontologies + 
operators. 
Is there a chace to 
merge the projects, they are separatedly very good but joined they will be 
greath,  
I may (if I can) 
help to manage this, with the permission of all of you! 
  
  
thanks 
  
 



  
  
    
      Andrés T. 
      Hohendahl	 
      CEO 
      – director	 
      	 
      (	 
      Office:+54 11 
      4711-9358	 
      )	 
      Mobile: +54 9 11 (15) 
      4148-1728	 
      *	 
      an...@pa...	 
      :	 
      www.pandorabox.com.ar