W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > February 2006

Re: GRDDL (split off from: Structured vs. Unstructured)

From: John Madden <john.madden@duke.edu>
Date: Tue, 14 Feb 2006 13:54:27 -0500
Message-Id: <D095A2B7-1641-4A38-AC29-DC04E83462F5@duke.edu>
To: HCLS <public-semweb-lifesci@w3.org>

Xiaoshu,


> I agree all your point except the "inseparably" part.  I think the  
> intension
> of GRDDL is to bridge, as oppose to merge - the world of HTML/XML  
> and RDF
> world. The former is intended for human and the latter for machine  
> (for
> instance for a better and precise web crawler to understand web  
> pages).  If
> the sole intent is to offer RDF description, a simple <link> tag  
> pointing to
> an RDF document will suffice. But using xslt transformation, it  
> saves the
> authors from "repeating" him/her-self.

Interesting point. I have to think this one over. I agree that a  
simple <link> to the finished RDF would work as well in many  
instances, and I agree with your point about avoiding repetition.

But if that's the sole advantage of GRDDL, I don't see it as a strong  
argument in favor of its use. That's why I've always interpreted  
GRDDL as something more, namely an early "web-of-trust" standard,  
where its use entailed some implicit contract with the rest of the  
web regarding semantic intent.

>
>> Question #2: Will this work for the case where the instance author
>> **doesn't** explicitly know the actual RDF triple set up
>> front, and the referenced extraction transform is actually
>> acting as a "language processor" to generate triples "that
>> thereby see the light for the first time"?
>
> I doubt a "yes" answer. SW technologies are designed for  
> representing rather
> than mining the knowledge. For example, someday when SW is matured  
> enough,
> you may be able trust your software agent with your credit card to  
> help you
> find and book your next flight to F2F meeting. I am not sure,  
> though, how
> much you can trust your agent with information mined from free text.
>

Actually, I completely agree with you here. I doubt it seriously too.  
That's not to exclude that some individual out there might  
voluntarily entrust his/her semantics to a machine, for some  
particular purpose. (For example, a semantic web researcher, who was  
doing a live demo project. Or even a doctor who had set up a well- 
tuned natural language processor to encode records for a clearly  
defined purpose.)

John
Received on Tuesday, 14 February 2006 18:55:05 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:00:42 GMT