Re: [dotNetRDF-Develop] Question about dotNetRdf library

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

Hi Andres

I know you emailed me the other month and I tried to reply but your email servers rejected my attempts at responding - you can see my original responses at http://sourceforge.net/mailarchive/message.php?msg_id=27140153

With regards to Euler the .Net implementation is not dead it is called EulerSharp - http://sourceforge.net/projects/eulersharp/ - and is still developed by Jos De Roo
The reasoners in dotNetRDF are certainly less powerful than Euler, reasoning has never been a huge priority for us.  As a quick way to integrate the two libraries you could use Euler to do the reasoning and use the output is produces (which I assume is in N3 format) and parse that into dotNetRDF to use as you wish.

It is probably possible to port the Euler reasoner onto dotNetRDF's API model but I have no idea how much work this would entail, if you'd like to do this and contribute it back to the project that'd be great.

Regards,

Rob Vesse

----------------------------------------
From: "Andres Hohendahl" <an...@pb...>
Sent: 01 April 2011 02:31
To: rv...@do...
Subject: Question about dotNetRdf library 






Hi Rob 
Vesse 
  
I see you bave don a 
wonderful job with all the stuff on dotnet.org, 
congratulations! 
Hope one day I can 
have a depp understanding of all the standards and formats you included 
inside! 
  

My purpose is to 
make a simple reasoner for a NLP dialog SDK system I am building for the las 5 
years,  
The idea is get a system packed with huge owl/rdf 
ontologies spread in the wild, so it may rason, link names find relations etc. 
 
Making a semantic 
web reasoner for linking into a simple human-machine dialog + reasoner 
system. 
  
Have to confess I 
have not much RDF-N3 etc. knowledge, better than the triple notation the 
rules-ontology hierarchy and some idea of reasoning. 
Now I have some 
questions to make.  First I will intoduce you the elements that confused 
me. 
After seeing your 
library I saw another similar library called EulerDotNet, created by Jos Roo, 
from AGFA and member of the W3C 
This library claims 
to be the fastest N3/DRF reasoner because it's using a Prolog+Euler reasoner 
using a ultra-fast java to prolog interface, all contained inside a big jar 
container. 
This project also 
has a .NET implementation, but it seems to be abandoned sometime between 2005 
and 2006. 
  
I tried to link all 
this together because the Euler Reasoner works really fast! and allows many 
problems (classical Socrates and Einstein ones) to be solved in a 
glimps! 
he problem is that 
people like yuo and me and many more, do not document all the stuff, specially 
3-rd party libs and code-blocks, so I am really confused on the compatibility of 
all this systems. I see a Prolog reasnones is necessary to solve FOL (First 
Order Logic) and a fast and intelligent algorithm like Euler's one may be 
good to be used with huve triplets and rules stores, to achive something in real 
time (not academoc work). 
  
So the main question 
is ¿did you see this libs? is the resoner you iimplemented as good as their's 
(sorry, I cannot inferr it from the docs) As my limited idea, it is not as good 
as Euler, and aldo  I found a C# native Prolog calles Yield Prolog, which 
is embeddable inside a library and works as fast as the best Prolog (they 
claim), and all this may be tied up together to make a bertter library, I 
guess 
  
I lay up the 
threads to this ideas, may be good, may be wrong, you'll see and have the last 
word 
  
I can help a little 
to tie them together with my limited capabilities  
(I am a C#, Java, 
Pascal, etc. programmer since 8 years using C# under .NET, written over 600k 
lines of code .. that is working...!). 
  
best 
regards. 
  
 



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