[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Semantic Web and First Order Logic

• To: xml-dev@l...
• Subject: Re: Semantic Web and First Order Logic
• From: "Thomas B. Passin" <tpassin@c...>
• Date: Thu, 20 Mar 2003 20:20:20 -0500
• References: <15725CF6AFE2F34DB8A5B4770B7334EEEAD0AB@h...>

[Bullard, Claude L (Len)]

> Is it true or false that the semantic web
> limits the use of First Order Logic?
>
> I realize that FOL has the undecidability
> problem, but is sufficient for everyday
> reasoning and is the most widely used
> logic in business.

I do not think that ordinary business applications today come close to using
full FOL.  Mostly things run on ordinary assertions (this would cover
relational databases), and sometimes on rulesets.  Ordinary markup languages
are EC - Existential Conjunctive - only one sector of FOL - and they are
almost always enough for these kinds of uses.

I do not see just what you mean that the semweb "limits the use of" FOL?  Do
you mean that the semweb must use a subset of FOL?  I think that will be so
as a practical matter - full FOL would rarely be seen in the wild, I would
think.  For one thing, it is so specialized and complex that one really
needs an experienced specialist, from what I have observed.

> One would think that
> the semantic web as a business application
> framework would use it profusely given that
> otherwise, the existing dominant base of
> relational business systems users would
> likely not want to be "on the semantic web".
>

I think almost everything people want to say on the semweb will be sayable
with EC, because most of it will be simple assertions of facts or
properties.  The ontology builders will want more, and OWL DL could be seen
as an effort to fill the bulk of that need.

There will be those who want to have less limitations that DL imposes, and
it remains to be seen how that can work on the open web without ending up
with mostly undecidable problems soaking up large processing times.

More likely would be some extensions to FOL (like modalities of various
kinds), but applied to a subset of FOL.  After all, you will probably want
to be able to make assertions about the reliability or probability or period
of validity, etc., of some assertion, and you need some extensions to do
that, AFAIK.

Cheers,

Tom P

• Follow-Ups:
  • Re: Semantic Web and First Order Logic
    • From: Miles Sabin <miles@m...>

• References:
  • Semantic Web and First Order Logic
    • From: "Bullard, Claude L (Len)" <clbullar@i...>

• Prev by Date: Re: Opinions
• Next by Date: Re: Semantic Web and First Order Logic
• Previous by thread: Re: Semantic Web and First Order Logic
• Next by thread: Re: Semantic Web and First Order Logic
• Index(es):
  • Date
  • Thread