I am sure there is a really good reason why none of the experts on this thread
have mentioned the "NULL' value in RDBMS practice, but for the life of me I
cannot guess it. It's simply not true that "not found" must mean "not exists"
in RDBMS (nor SQL) -- NULL can be set as an explicit or default value ... the
meaning of which is more akin to "not yet specified (and therefore not yet
known, one way or the other)". That seems more OWA than CWA, if I am following
you good people correctly. (01)
Cheers,
BobN (02)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F Sowa
Sent: Tuesday, January 22, 2013 2:35 AM
To: ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum] Knowledge graphs by Google and Facebook (03)
William and Kingsley, (04)
I'll be traveling for the rest of this week, so this is my last note
on this thread -- at least until next week. (05)
WF
> So, this is what I think I have learned from this discussion, that I can
> apply:
>
> "The only known semantics that makes SQL and many other computer-usable
> (for the things we want computers to do) -languages work is something
> called the "Closed World Assumption". This so-called "assumption" is
> in fact obviously false, so when we go about using the the results of
> queries in our programs and interpreting them as people, we need to be
> very careful to take this weakness of the computer languages into account. (06)
Actually, the theoretical issues are more subtle, and the practical
issues are simpler. (07)
The CWA is the assumption that makes SQL, Prolog, and many other
systems amenable to classical semantics for first-order logic.
Another name for "classical" is "Tarski-style", since Alfred Tarski
stated the method, called model-theoretic semantics, around 1930. (08)
There is also more recent work on nonmonotonic reasoning, which
deals with the Open World Assumption (OWA) and many related issues,
including default logic, negation as failure, circumscription,
and belief revision. There is a large "cottage industry" for
producing papers and dissertations about these issues. (09)
The simple solution for practical purposes is the one you suggested:
just recognize that nearly all databases are incomplete (in the
sense that much of the expected information is missing). Then
remember that 'not' in the query language means 'not found'. (010)
As long as you remember those principles and you write suitable
exception handlers for dealing with the inevitable failures that
arise, your systems will work reasonably well for most practical
purposes. When they fail, you can do what Microsoft does -- issue
patches every Tuesday. (011)
KI
> A DB2 relational dbms that in version 10 is enhanced with RDF data
> model storage and SPARQL query language support (012)
Yes. Oracle also handles RDF and SPARQL. The vendors who sell graph
based systems support SQL, and the relational vendors support SPARQL. (013)
The conceptual schema proposals developed by the DB community in the
1970s (and '80s and '90s) were intended to support interoperability.
They tried to make a clean separation between the logic and the
implementation details. The major obstacle was the vendors who
didn't want interoperability with products sold by their competitors (014)
John (015)
_________________________________________________________________
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 (016)
_________________________________________________________________
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 (017)
|