RE: What is truth anyways? was: [...]

At 9:44 AM +0100 6/12/02, Peter Crowther wrote:
>  > From: Jim Hendler [mailto:hendler@cs.umd.edu]
>[...]
>>  Interesting Pat, so you're saying that when I stick my little plastic
>>  card into the Automated teller in Italy, and it hands me Euros
>>  charging an appropriate exchange rate against my machine in the US,
>>  that they are using a formal model theory to make it work -- can you
>>  show it to me??    err, perhaps sometimes you underestimate what can
>>  be done with "social agreements" instead of pure logic...
>
>In this case, those 'social agreements' are specs of various banking
>interchange formats plus places to download exchange rates and maps of card
>number prefixes to issuers.  These are all written by humans, interpreted by
>humans, and turned into (often buggy) special-purpose code and text files by
>humans.

ahh, so you agree with me, we can make this stuff work more than well 
enough to trust without having the formal semantics!

>
>The snag is that, with the SW, we're trying to remove this human
>interpretation and special-purpose code writing.

what a weird idea - I'm trying to make it significantly easier and 
faster and cheaper for humans to do these things, that I agree, and 
to make more of the information explicit, that I agree - but to 
replace humans in the design and development of these things - that's 
a dream even a crazy like me thinks is impossible.

>  That, to me, is what makes
>the SW interesting: that I can create (essentially) some fancy data that
>complies with some standards, and be certain that I can convey its formal
>meaning to any other SW agent that is compliant with those standards.

sure - but it ain't turtles all the way down - sooner or later some 
human is implementing something, and that has some bugs and the 
semantics helps us find and debug them sometimes -- but a lot of 
other folks are creating logics and things "wrong" and if we can't 
live with it, we're dead.  The notion that making things formal takes 
away the social issues in agreement is wrong -- want a great paper on 
that?  The late Alan Perlis (early Turing award recipient) wrote a 
famous paper about mathematical proofs as social processes in the mid 
1970s.  [1] is a set of easy "epigrams" that appeared in  ACM 
sigplan, 1982 (example: "It is easier to write an incorrect program 
than understand a correct one.") - I cannot find the text of that 
paper on the line - but the citation is

De Millo, Richard, Richard Lipton, Alan Perlis. "Social Processes and 
Proofs of Theorems and Programs." In New Directions in the Philosophy 
of Mathematics: An Anthology: Revised and Expanded Edition, edited by 
Thomas Tymoczko. Princeton, NJ: Princeton University Press (1998) 
267-285.

I strongly recommend finding it - it is far more eloquent than I am 
and the argument is made by three of the most famous mathematical 
computer scientists of its day.

>  In
>the case of some financial transactions, the standards-compliant data
>represents financial transactions.  That's boring.  I'm interested in
>transferring as much formal meaning as possible around the Semantic Web and
>in ensuring that meaning can be interpreted --- unambiguously and according
>to my original intentions --- by the widest range of automated systems.

I agree with everything in this quote except the way you use the word 
formal -- but then, I suspect that's a philosophy of life issue more 
than a truly technical one - c.f. Bob Abelson's famous quote about 
neats vs. scruffies.

>Natural language and comments *in the data*, as opposed to the specification
>of the data format, are not going to help me achieve this, as there are no
>humans in the loop to interpret that natural language and modify the
>automated systems to understand that data.  The natural language comments
>are not part of any social agreement; they are merely included by one party.

sorry, I wasn't arguing against that - while I think NL comments are 
very useful for implementors, I agree theyr'e not very useful for 
implementations.  I was arguing against Pat's axiom, not his 
conclusion...
  -JH


[1] http://www.cs.yale.edu/homes/perlis-alan/quotes.html


-- 
Professor James Hendler				  hendler@cs.umd.edu
Director, Semantic Web and Agent Technologies	  301-405-2696
Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742	  240-731-3822 (Cell)
http://www.cs.umd.edu/users/hendler

Received on Wednesday, 12 June 2002 05:56:08 UTC