Re: Comment on "Meaning and the Semantic Web"

> > > These are shock-jock type of arguments
> >
> > Huh?
> >
                                                                                                             
> Shock jocks make their ideas feel profoundly important by
> attaching other highly emotionally charged (shocking) ideas to
> them. 

Our ideas are, doubtless, profoundly pedestrian. But the issue we're
dealing with (communication using next-generation Web technology) 
is a profoundly important one. That importance is sometimes 
illustrated by connecting it with things that matter to people's 
every day lives. Like being happy, sad, rich, poor, oppressed, 
misrepresented, misunderstood. 

>	The examples in Dan's and Jeremy's emails have that feel to
> me - anti-homosexual bias, racism, but so does totalitarianism.

I chose emotive examples on purpose: it was the point of the 
exercise, even. All the while that everyone agrees about everything, 
we can be utterly naive about RDF. All the while we agree to disagree
by disputing the facts but completely sharing the vocabularies within 
which those facts are expressed, we have it pretty easy too. The
problems come when the terms used to express claims about the world are 
disputed terms. It shouldn't be suprising that terms are more contested 
when they describe areas of disagreement and emotive debate. 

I could have picked less emotive terms (eg. relating to definitions of 
employment, unemployment) which are nevertheless politically charged and
contested. I could also have picked more emotionally charged topics and 
terms, eg. notions of alive/living/etc in the context of debates around
abortion policy. Or the naive definitions of 'terrorist' we see floating around 
in DAML+OIL or OWL, as if encoding in a formal language 
automatically brought with it credibility and intellectual rigour.

Or perhaps we could be discussing the definition of "torture" as 
"an act by which severe pain or suffering, whether physical or mental,
is intentionally inflicted on a person"[1], investigating whether 
organisations which don't buy into a common definition are 
nevertheless capable of communicating using RDF/OWL technology. One 
person's torture is another's "rough handling", I'm told. I
preferred to stick with the all-time classic topic of Sex. There are 
vastly more topical and controversial themes, very relevant to our work,
that we could have focussed on instead.

We may be discussing emotive issues, but it is the corner cases that test the
technology; I don't think we can stick to areas where things are
controversy-free without missing the point. For areas of discussion where 
terminology isn't politicised, RDF/OWL pretty much works out of the box. 
For the rest, we still have some work to do. Considering such cases doesn't make us 
'shock jocks', though given the emotional baggage that comes with the 
territory, it is probably worth reminding folks to be extra careful with
their examples.

The Web is the world's most important system for publishing and sharing 
documents and data, and it needs to work for everyone, including parties
who are mutually antagonistic. RDF/OWL needs to be deployable in a world
where language and terminology is politicised, because that's the kind
of world we live in.

Incidentally, the RDF design inherits from PICS, W3C's earlier 
"Platform for Internet Content Selection". The 1996 note on 
"Using PICS Well"[2] is worth bearing in mind when we think of the
social impact of our technical work. The PICS design decentralised the
creation of content rating vocabularies in part through a recognition 
that rating and classification schemes can embody contested,
controversial worldviews. RDF's design differs from PICS' in that a
single instance document can now draw upon terms defined in several 
independently created vocabularies. By making this possible, we don't 
necessarily make it wise. (I argued a variant on this in [3], dunno if 
slides are any use on their own though). My understanding of what 
we're trying to do here (public-sw-meaning) is that we're probing at the 
different social and technical notions of 'shared meaning' that get 
tangled up when such systems are deployed in the wild.

cheers,

Dan



[1] http://www.amnesty.org.uk/torture/definition.shtml citing 
the UN Convention against Torture and Other Cruel, Inhuman or Degrading
Treatment or Punishment, http://www.unhchr.ch/html/menu3/b/h_cat39.htm

[2] http://www.w3.org/TR/NOTE-PICS-Statement

[3] http://www.w3.org/2001/sw/Europe/talks/20040309-danbri/

Received on Friday, 4 June 2004 15:00:11 UTC