[Fwd: Re: EE Times - Sony lab tips 'emergent semantics' to make sense of Web]

["Davis, Peter" <peter.davis@neustar.biz>]

of overlapping interests with our ontology work... link to the EE
article is at the bottom.

--- peterd

-----Forwarded Message-----
> From: Frank Manola <fmanola@acm.org>
> To: Graham Klyne <GK@ninebynine.org>
> Cc: Arjohn Kampman <arjohn.kampman@aduna.biz>, www-rdf-interest@w3.org
> Subject: Re: EE Times - Sony lab tips 'emergent semantics' to make
> sense of Web
> Date: Fri, 05 Nov 2004 09:39:45 -0500
> 
> 
> 
> 
> A couple of comments
> 
> Graham Klyne wrote:
> > 
> > I saw this (or a version of it), and had two leading thoughts:
> > 
> > 1. The new technology was presented as a competitor to the Semantic
> Web 
> > technologies, in a way which suggests to me a mistaken (or, at
> least, 
> > very narrow) view of what Semantic Web technology is about.  To my
> view, 
> > a system that can perform automatic ontology extraction/generation
> is a 
> > natural complement to the W3C raft of technologies (if it works).  
> > There's nothing in OWL and RDF, for example, that demands that 
> > application developers do ontological markup by hand, or that such 
> > information be embedded into all data sources.  I quite frequently 
> > stumble across projects to interpret existing data sources/formats
> as 
> > RDF (e.g. the calendaring task force work to interpret iCalendar
> data,  
> > various proposals to present relational database information as RDF,
> etc.).
> 
> I think this is the correct view.  Part of the problem is that lots of
> discussions of the Semantic Web (and its "competitors") cast this as 
> being somehow all one technique vs. all of another.  IMO, the Semantic
> Web, per se, doesn't really care how the semantic information is 
> acquired, just that it's there, and is available to programs that want
> to process Web information.  All available sources of such information
> clearly need to be exploited, which includes emergent semantics. 
> However, this information also includes definitions that already exist
> in database schemas, industry nomenclatures of various sorts, and so
> on 
> (and emergent semantics might help in amplifying such information as 
> well).  One of the things that such articles indicate is that there's 
> clearly a lot more work needed to accurately convey what the Semantic 
> Web is about to a wider community!
> 
> > 
> > 2. Sony's new technology has apparently just been patented.  Which I
> > think kills it dead in the water as a potential replacement for RDF,
> OWL 
> > and friends.  The level of fundamental infrastructure needs to be 
> > completely free and open to survive as such, IMO.
> 
> This sort of depends on what the patented technology actually does. 
> For 
> example, it seems to me that any semantics that emerges from 
> conversations between agents needs to be captured in some concrete
> form, 
> so that it doesn't have to be rederived for each conversation, and can
> be used as a building block for further communication.  The concrete 
> form might very well be OWL, or some amplification thereof.  The 
> detailed processes by which the semantics derivation is performed
> might, 
> on the other hand, be patentable.
> 
> --Frank
> 
> > 
> > #g
> > -- 
> > 
> > At 10:23 05/11/04 +0100, Arjohn Kampman wrote:
> > 
> >> Someone just notified me of the following article on EE Times:
> >>
> >> http://www.eetimes.com/article/showArticle.jhtml?articleID=51201131
> >>
> >> "As the World Wide Web Consortium hammers out specifications on how
> to
> >> recode the databases of the world so that natural-language queries
> can
> >> be intelligently answered online, Sony Corp. says it has found a
> better
> >> way."
> >>
> >> Don't know what to think of this article. Is it for real? Comments,
> >> anyone?
> >>
> >> Arjohn
> >>
> >> -- 
> >> arjohn.kampman@aduna.biz
> >> Aduna BV - http://aduna.biz/
> >> Prinses Julianaplein 14-b, 3817 CS Amersfoort, The Netherlands
> >> tel. +31-(0)33-4659987  fax. +31-(0)33-4659987
> > 
> > 
> > ------------
> > Graham Klyne
> > For email:
> > http://www.ninebynine.org/#Contact
> > 
> > 
> > 
>