Re: Decidability
Jay,
Thanks for sending us those references, which make some
useful observations. I certainly agree that the issues
are not "simple". Even when there are clear-cut answers
to specific questions, it's not clear whether those are
the right questions to ask. I like the following remark
by James Spall:
> Algorithms should be implemented in context: “Better a
> rough answer to the right question than an exact answer
> to the wrong one” (Lord Kelvin)
Source: http://www.jhuapl.edu/ISSO/PDF-txt/Chapter-01.ppt
I would also qualify the following claim:
> In the past ten years, the development of the world-wide
> web has caused a revolution in the world of information
> processing. The framework for the publication and exchange
> of scientific results is changing deeply and rapidly.
The WWW marks the start of widespread public awareness, but
for most of us in the comp. sci. community, the real revolution
began with Arpanet and similar networks (for me, it was IBM's
VNET) in the 1970s. HTML certainly formatted the files in
a prettier form, but we were using email, instant messaging,
file transfers, and pretty formatting with GML and LaTeX
twenty years before the WWW.
I would compare the WWW "revolution" to the PC "revolution":
nothing changed in theory, but there was an enormous shift
in availablity to a wider audience. That's important for
practical reasons, but not in principle.
> Today, a large part of scientific knowledge is available
> not only in symbolic form (i.e., in texts), but also
> in a format which supports machine-based search, analysis,
> and composition.
That is Semantic Web hype. The raw ASCII text files are
still the major input to all the search and analysis engines.
Google deliberately ignores the so-called "metadata" because
it is highly unreliable and frequently abused by spammers.
> It is rather clear that new kinds of "inference" and
> "composition of propositions" have to be developed to
> handle the potentials of the web adequately.
More hype. There is nothing in the current R & D for
the Semantic Web that begins to approach the level of
research in artificial intelligence and computational
linguistics of 20 years ago. I'm not claiming that
the R & D of the 1980s is adequate -- but merely that
translating it into XML won't solve any real problems.
> Sometimes I have the impression that we are living
> in a golden age of logic but that logic does not
> know it.
The real golden age of logic was the 13th century. That's
when Peter of Spain published his "Summulae Logicales",
which was the principal textbook on logic for the next
300 years. Even until the early 20th century, most
textbooks on logic were watered down versions of Peter's
book. I still recommend it to anybody who wants to learn
the fundamental principles of logic and ontology. For a
good English version, see
Joseph, Sister Miriam (1937) _The Trivium: The Liberal
Arts of Logic, Grammar, and Rhetoric_, Third edition 1948,
reprinted by Paul Dry Books, 2002. ($11.87 in paperback)
This is what college freshmen were studying for over 700 years
(1239 to 1948). What is truly sad is that very few university
graduates today learn that level of logic. Most of those who
are exposed to "modern" symbolic logic hate it, and only a
tiny fraction really learn it.
And by the way, Peter of Spain later became Pope John XXI.
Can you imagine being a logician as a promising career path
for any aspiring papal candidate today?
John Sowa