Actually, I've just thought of a small snag in that beautiful picture, which
again concerns negotiations between ontologies -- translations between
languages. It's not difficult to see that this might apply across ontologies
defined in the same language.
There's the classic example in John F. Sowa's "Knowledge Representation"
book on p.410 concerning the equivalences between the English
'river' and 'stream'
and the French
'fleuve', 'riviere' and big 'riviere'
'River' in English maps to both 'fleuve' and big 'riviere' depending on
context.
And 'riviere' in French can map to 'stream' in English, but can also mean
big 'riviere', again depending on context.
This is because in French a 'fleuve' is a river that flows into the sea, and
a 'riviere' is either a stream or a river that flows into another river.
This is why I worry about simple label PSIs.
But I also think it complicates the notion of basin of attraction
significantly to something more like contextual overlap.
And that overlap might only be partial at the time of comparison.
Time to go read that big book on similarity and analogy I have in the
pending pile.
Cheers,
Peter
----- Original Message -----
From: "Jack Park" <jackpark@thinkalong.com>
To: <unrev-II@yahoogroups.com>
Sent: Saturday, September 08, 2001 6:15 PM
Subject: Re: [unrev-II] Semantic Community Web Portal
> Hi Peter,
>
> Elsewhere, I have argued that there should not be "point" attractors as
> binding points, but rather "attractor basins" as binding regions. You
have
> stated that case from a powerful point of view.
>
> The question, as I see it is this: how to implement attractor basins?
>
> My conjecture has been that the PSI that Bernard mentions and discusses
> quite eloquently in his chapter in the forthcoming book on XML Topic Maps
> should be something the equivalent of an entry into a tiny DKR for the
> concept it represents. That DKR is, itself, an evolving space of
> contextual definitions, discussions, and representations, rather than a
> lone "point in space" representing for all humankind a lone "binding
point"
> for the concept to which it provides reference.
>
> That DKR is, itself, the place where ongoing, never-ending, always
evolving
> negotiation occurs. In the strongest sense of facilitated evolution, a
> tiny DKR for each "binding point"-->"binding region" is just one
> implementation of the requirement that each concept in ontology space be
> capable of being negotiated.
>
> Cheers
> Jack
>
> At 05:24 PM 9/8/2001 +0100, you wrote:
> >Bernard Vatant wrote:
> > >Bottom line :
> > >Binding separately developed but overlapping ontologies will need
> > >non-ambiguous stable binding points. Topic Maps people call that
Published
> > > Subject Indicators.
> >
> >Dear Bernard,
> >
> >(At the risk of being grievously outspoken, opinionated, re-opening old
> >wounds, and so forth...)
> >
> >Not necessarily. It does require *negotiation* of degree of equivalence,
or
> >measure of similarity of concepts for the purposes of that particular
> >agreement to communicate.
> >
> >There are aslo deep, future, ethico-political issues tucked away in that
> >assertion of need, on the assumption that Topic Maps will become/are
> >becoming a significant technology in global human affairs. (And from what
I
> >know of the process, everyone is so beloved of mad-dash development of
the
> >technology that I am uncertain as to whether anyone in that group could
> >possibly have had the time to consider those issues fully.)
> >
> ><rant severity="Gee, I wish I didn't have to...">
> >I don't agree with cultural imperialism;
> > I do believe that *simple* mechanisms of cultural compromise can often
> >result in losses for both parties;
> > I don't believe that relinquishing the idea of concepts as theories is
a
> >good idea long term, particularly in respect of innovation;
> > I would love to see someone nail down the PSI for 'vagueness' first,
so
> >that we can all point at it together with a percentage indicator when
none
> >of us agree completely about any other PSI we're pointing at later on.
> >How do I know that a particular PSI I've created because I couldn't find
any
> >other PSI that expressed what I needed to have someone else agree about
will
> >be construed in exactly the same manner by anyone else?
> >Will I need a Topic Map of the PSIs so I can find a PSI on the Web that
> >suits my purposes?
> >Will I need a Topic Map of Topic Maps of PSIs...?
> ></rant>
> >
> >I know the technology well, I know its advantages and its limitations.
> >
> >I know that in most senses Topic Maps is still a prototypical technology,
> >and that therefore I am not asking that they be the complete solution to
> >Life, The Universe, and Everything, all at once, now.
> >However, I would respectfully request of the Topic Maps standard
development
> >community that they look very hard at the question, "Is there a better
way
> >of doing X?" all the way down the line, not just for the sake of more
> >efficient algorithms, and not rush things.
> >
> >I want these things to improve the quality of users' lives.
> >Faith without due consideration is dangerous.
> >
> >Best regards,
> >Peter
>
>
>
> Community email addresses:
> Post message: unrev-II@onelist.com
> Subscribe: unrev-II-subscribe@onelist.com
> Unsubscribe: unrev-II-unsubscribe@onelist.com
> List owner: unrev-II-owner@onelist.com
>
> Shortcut URL to this page:
> http://www.onelist.com/community/unrev-II
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
>
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Do you need to encrypt all your online transactions? Secure corporate intranets? Authenticate your Web sites? Whatever
security your site needs, you'll find the perfect solution here!
http://us.click.yahoo.com/wOMkGD/Q56CAA/yigFAA/IHFolB/TM
---------------------------------------------------------------------~->
Community email addresses:
Post message: unrev-II@onelist.com
Subscribe: unrev-II-subscribe@onelist.com
Unsubscribe: unrev-II-unsubscribe@onelist.com
List owner: unrev-II-owner@onelist.com
Shortcut URL to this page:
http://www.onelist.com/community/unrev-II
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
This archive was generated by hypermail 2.0.0 : Sat Sep 08 2001 - 12:58:05 PDT