[foaf-dev] Group vs Class

Peter Williams pwilliams at rapattoni.com
Sat Jun 12 06:06:45 CEST 2010


What I liked about the foaf spec (once I got passed the academic barrier) was that it exhibited what it was teaching: that the system of types (deduced only after reasoning) told the same story as did the property/class tags that  codified a domain language representing intuitions about the same concept set. The tags were a facet of the type system, that is -  easier to remember and work with than the underlying types. The compiler (reasoner) would work with the underlying types, when "computing". If one as a modeler bothers to think like a compiler, there would however still a natural domain-centric feeling the knowledge being expressed in that form (but one needs a certain self-confidence and training to think of knowledge systems in terms of types and behaviours, rather than relations/properties between entities). 

And this is what frustrates me about semweb in general.

Folks model applied domains using semantic constructs, but then seem to deny the applicability of the reasoning engines and the role of non-http-native agents (because of all their complexities in adoption). I typically find myself learning from the modeling excellence, but then fall back on using good old fashioned non-semweb tools instead (because of the denial that the higher tools and agent-based protocols are "truly" relevant).

Its almost like the movement sabotages its own adoption, so it can stay a research-phase project.

Just a peon's view.

-----Original Message-----
From: foaf-dev-bounces at lists.foaf-project.org [mailto:foaf-dev-bounces at lists.foaf-project.org] On Behalf Of Dan Brickley
Sent: Friday, June 11, 2010 12:42 PM
To: Ian Davis
Cc: Dave Reynolds; foaf-dev
Subject: Re: [foaf-dev] Group vs Class

On Fri, Jun 11, 2010 at 6:07 PM, Ian Davis <me at iandavis.com> wrote:
> Hi Dan,
>
> I was hoping you would have thoughts on this :)

Thoughts are easy; figuring out where to take the design, and how to
decide - a bit harder!




More information about the foaf-dev mailing list