RE: Semantic Suggestions please ... for oilit journal

>> hackers would convert his oil-IT journal's semistructure data to
Django/Rails/ARC fields in an afternoon, set up some Sphinx searching, and
move onto their next gig without even uttering "Lucene"

>Let's not divide the world between Hackers and Consultants please... 
Plenty of companies and universities give away vast and precious piles 
of work (code) for free. Even large universities. Generalisations at 
this level don't help anyone, really.


Just catching up on this thread after a short break for the 'daytime job'.
Re the hackers vs consultants - I was actually hoping to hack something
together myself. I didn't think that this would turn out to be quite so
rocket-sciency. 

Lets try another tack.... 

In so far as a semantic version of my website will be amenable to trawling
by semantic agents of some sort or another, what they be looking for? 
A text based news website might look like an RSS feed (that's a bit
semantic) with some embedded controlled vocabularies. Say it's a fruit
newsletter - with a vocabulary of companies - bigFruitCo, LittleFruitCo and
another vocabulary - apple, banana, orange. So now when I write my article 
"I was tasting some great BANANAS from BIGFRUITCO the other day ...." the
capitalized words have some code behind them that says BANANAS are a the
fruit in the list on my website - where presumably I can add lots more
information about who owns the company etc. Maybe even "BigFruitCo bought a
20% share in LittleFruit" co on such and such a date ... which could be
structure or text - but is definitely semantic - but I digress. 

My main questions are "where code is an agent looking for in the controlled
lists?" "Is there any standard way of putting these in a file?" A code
sample or pointer to such would be nice.

I suppose that I am looking to be able to deploy a homebrew kind of
OpenCalais on my website. The OpenCalais EULA did seem rather intimidating -
not that I actually read it...

Neil

Received on Thursday, 30 October 2008 14:42:59 UTC