Kingsley, Simon, and Pavithra, (01)
One of the goals I have been trying to promote for over 30 years is
the development of a unified framework and a systematic methodology
for semantic systems. That was the goal of my 1984 book, _Conceptual
Structures: Information Processing in Mind and Machine_, and my 2000
book, _Knowledge Representation: Logical, Philosophical, and
Computational Foundations_. (02)
Over the years, I participated in a variety of interdisciplinary
conferences and organizations that were working to promote such
integration. As an example, following is the proceedings of a
Workshop on Data Abstraction, Databases and Conceptual Modeling,
at Pingree Park, Colorado, in 1980: (03)
http://www.sigmod.org/publications/dblp/db/conf/sigmod/pingree80.html (04)
That workshop brought together a good group of experts in AI, database
systems, and programming languages, who presented a stimulating
selection of ideas. But rereading the proceedings is depressing:
it shows how little progress has been made in the past 30 years.
You could take any one of those papers, dress it up with a few new
references, throw around some words about ontology and the Semantic
Web, and present it at a similar conference today. (05)
But for the same reason, that workshop is encouraging. It shows that
the old guys from Aristotle to Wittgenstein had a pretty good coverage
of the basic issues in semantics. The most important new things that
have to be said and done are about the technology for implementing it.
That was the point I emphasized in the following slides: (06)
http://www.jfsowa.com/talks/iss.pdf
Integrating Semantic Systems (07)
The most important progress in the past 30 years has been in the
technology for implementing those ideas. We have better tools and
algorithms today, much bigger, faster, and cheaper computers to run
them on, and a much bigger and faster descendant of the old Arpanet. (08)
Some of the projects that I surveyed in the iss.pdf slides have taken
advantage of that technology, and they have demonstrated feasibility.
But what I find frustrating is that the mainstream, as illustrated
by the components of the Semantic Web, is still running the algorithms
of the 1980s on the hardware of 2010. There is a lot of whiz-bang
cosmetics in flashy GUIs, but no semantics beyond what was implemented
in research systems of the 1980s. (09)
KI:
> I agree with the general "skunk works" theme 100%, but Java isn't a
> great example today IMHO. Lots of bloat in Java land (codebase and
> community process). (010)
I agree that the developments on top of Java include a lot of bloat.
But I was talking about the development process at Sun from OAK in
1990 to the announcement of Java in 1995. That development group
was similar to Lockheed's famous "skunk works". (011)
SS:
> But see Mettler et. al (2010), etc, for a different approach to secure
> java programming. I love the word fabric; I like to contrast fabrics
> with infrastructure, in that a fabric is much more comfortable against
> the skin, and fits itself to the contours of the individual. (012)
Yes. The fabric vs. web metaphor was the main reason for mentioning
that project. But I also liked the fact that they have been doing
a lot of R & D with a small group that has been trying to design a
more systematic framework than just a web of links. (013)
SS:
> A justification for W3C was given at ECHT94, online at
> http://www.w3.org/Talks/ECHT94-Keynote/Keynote.html (014)
Yes, but the conclusion is too vague: (015)
TBL:
> Things to take away
>
> * We have an information space
> * We have a way forward
> * We are in it together
> * It is very important to keep it open (016)
I can't quarrel with those points as guidelines for a committee.
But you can't do good design in a committee. It would have been
much better to promote design competitions among multiple groups. (017)
The practice of letting a large committee make "recommendations"
about untested designs is what destroyed the integrity of OS/360
-- as Fred Brooks so eloquently explained. (018)
SS:
> Nobody really was really paying much attention to SemWeb until
> Berners-Lee et. a (2001). It was difficult enough making the
> Web of Documents work. (019)
Nobody paid any attention to OAK-JAVA from 1990 to 1995. The
whole idea of the skunk works is to keep the plethora of chefs
and committee members from stirring the pot before it's ready. (020)
JFS:
> As a result, the only consensus on an architectural document was
> the familiar layer cake, which emphasized syntax over semantics. (021)
SS:
> The cake is a lie! Ask Jim Hendler! (022)
I've talked with Jim off and on for the past 20 years or so.
We sometimes agree and sometimes agree to disagree. (023)
PK:
> Linked Data as "a term used to describe a recommended best
> practice for exposing, sharing, and connecting pieces of data,
> information, and knowledge on the Semantic Web using URIs and RDF." (024)
I agree that's what the term is used for. And just doing that
is not bad. But we can do vastly, vastly more than that. (025)
What I would have liked to see is a series of independent design
competitions. For example, (026)
- Jim H. & Co. at Maryland could have extended SHOE. (027)
- Guha could have developed RDF with collaboration from DAML
and OIL. (028)
- Cyc could have developed a multi-tiered semantic system
with a simple entry level that led up to full-scale Cyc. (029)
- Many groups had excellent rule-based systems integrated
with relational and/or object-oriented DBs. (030)
- Development methodologies based on UML diagrams could have
integrated mainstream IT with the WWW for a much smoother
and less painful transition. (031)
Those are just five examples, but an open-ended variety of other
systems could have competed. I would love to see what would come
out of a multi-stage design competition with the option of letting
any group borrow ideas from the others at each stage. (032)
I'm sure it would be far superior to what the SemWeb has become. (033)
John (034)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (035)
|