Re: [ontolog-forum] Amazon vs. IBM: Big Blue meets match in battle for t

[Kingsley Idehen <kidehen@xxxxxxxxxxxxxx>]

On 7/26/13 8:27 AM, David Eddy wrote:

Kingsley -

On Jul 25, 2013, at 8:16 PM, Kingsley Idehen wrote:

[1] http://bit.ly/Y6TIfs -- Using SPARQL to Integrate Disparate Data via InverseFunctionalProperty (IFP) relations
[2] http://bit.ly/WmKlJ0 -- Demonstrating Reasoning via SPARQL & SPARQL 1.1 Property Paths .

To my needs these examples do not solve the challenge of understanding that M0760 and MENSA-FL (100% real data names) are actually synonyms.

"Ask the expert" is not a valid response.  Expert not available.  I'm on my own.

#1 - nice that they match, but it seems to depend on both items containing your email.  I seriously doubt I'm going to find someone's email embedded in a data structure.

#2 - some things in life are nicely hierarchical.  A lot of things are not... particularly the use of language, and even worse for the use of technical jargon.

To repeat: in the context of legacy systems (definition of a legacy system is a system that works, produces reasonably predictable results & people depend on... regardless if it's a day old or 40+ years old), is there a SW mechanism that helps analysts/programmers understand what the labels in these systems mean?

A system written in R & Hadoop (assuming it's in production) is just as legacy as something written in Jovial.

___________________
David Eddy
deddy@xxxxxxxxxxxxx

To ease burden of thread tracking, I am going to repeat the response I just posted to the other thread (between yourself and Michael):

Note that there is an NLP and machine-learning dimension to all of this too. By that I mean: you can also integrate NLP services and machine-learning into the processing pipelines of Semantic Web stack services, as exemplified by Virtuoso's Linked Data middleware layer [1].


Many of the aforementioned NLP services [2][3] are customizable which is also where domain experts can make domain-specific tweaks. This is what you need for hooking legacy systems into these kinds of solutions, when dealing with blurb.

Beyond blurb, when the structured data representation is fine-grained (rather than coarse-grained) you can simply customize or use an existing provider/mapper/cartridge for translation to RDF based structured data representation.

The puzzle pieces are in place. The architecture of the World Wide Web (and its fine-grained semantic variant) is inherently loosely coupled. We just need to do a better job or building solutions that amplify the loosely coupled nature of the aforementioned platform and architecture. For instance, RDF zealotry ultimately creates the illusion of a monolithic silo which is utterly antithetical to the entire Semantic Web mission and vision.


Links:

[1] http://linkeddata.uriburner.com/sponger_architecture.vsp#how_it_works -- animated walk-through (note: meta cartridges are what hook into machine-learning driven NLP services from 3rd parties e.g., Spaziodatti, OpenCalais, DBpedia Spotlight, Zemanta, Alchemy API, and others).

[2] http://nlp2rdf.org/about -- nlp2rdf.

[3] http://spaziodati.eu/datatxt-entity-extraction-linking-api/ -- Spaziodati .

-- 

Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software     
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen

smime.p7s
Description: S/MIME Cryptographic Signature

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (01)