W3C home > Mailing lists > Public > public-rule-workshop-discuss@w3.org > August 2005

Re: Comments on * DRAFT * Rules Working Group Charter $Revision: 1.60 $

From: Christian de Sainte Marie <csma@ilog.fr>
Date: Tue, 23 Aug 2005 19:55:55 +0200
Message-ID: <430B632B.1030308@ilog.fr>
To: Dieter Fensel <dieter.fensel@deri.org>
CC: public-rule-workshop-discuss@w3.org

Dieter Fensel wrote:
> Comments on * DRAFT * Rules Working Group Charter $Revision: 1.60 
>  [...]
> Given the current draft, we would rather prefer not seeing a working 
> group or renaming the working group to "A working group on extensions of 
> OWL towards a not well-justified sub fragment of first order logic with 
> equality." However, with OWL-Full we already have a zombie in our 
> basement, that is, there is no need to create a second one?

Hi Dieter; I see that you are not happy with Sandro's draft :-)

Well, I can see your point. Actually, I agree with most of your
statements... but I believe that you are dead wrong nonetheless!

Actually, Jim Hendler said it all: as I understand it (and as it says
explicitely, as Sandro stressed), the charter is about specifying a
common notation to exchange rules between applications (each using their
own, different rule languages).

Actually, you raise the right point (and draw the wrong conclusion :-)
in your subsequent message:
> I expected W3C to define a charter for a rule language whereas the current
> charter is from my point of view either
>          - technically wrong, since it does not choose a rule language paradigm

Indeed, I think that it is a very important feature of the charter that
it does not choose a rule language paradigm. I do not know if this is
technically wrong, but it seems pretty right to me, from a practical
point of view.

The point is that you do define a standard based on what people do and
how they do it, not the reverse. In other words, if W3C specifies a rule
language (for the Web or else), it will most probably end up just being
yet another rule language (plus, specifying rule languages is not
especially a W3C competence, is it?); whereas, if it specifies a
(extensible) standard format or protocol for exchanging rules between
different rule languages, there it brings a real added value to all the
(existing and future) rule languages (plus, this is really where the W3C
competence is, I believe).

The exchange format should be as neutral as possible with respect to the
various logics, paradigms and languages, I believe; and applications, of
course. Then, a rule-based application would be expected to be able to
recognize whether it can do something with retrieved rules (or what it
can do with them) when translating them back from the exchange format
into its own rule language, either because the rules will explicitely
belong to a defined profile (and a specific language should know which
profiles it supports) or by examining the rules themselves (e.g. a
resolution-based engine would easily recognize if the rules are Horn
clauses) [1].

Hence the proposal to use FOL as its core. As Sandro wrote (in reply to
Michael Kiefer):
>> An exchange language is about XML syntax. You don't need semantics for
>> that---only an encoding.
> If the XML syntax doesn't have an associated semantics, how can you
> tell if you've mapped your language to the standard format correctly?
> Not everthing call "implies" means the same thing, right?

Actually, this is probably where the draft charter is misleading.
Instead of saying that "the core of the language will be FOL", it could
say that the core format will be a notation for FOL with (e.g.)
Tarski-like semantics. The choice of FOL is because most practical rule 
languages can be mapped onto a FOL notation, or so I suppose. Whether 
FOL with equality or leaving that to oracles, I do not know: if we are 
talking about an exchange notation, here, does it really make a 
difference? I mean, it depends on how the language you translate the 
rules into to use them handles that, or am I naive, here?


[1] Btw, one thing that is missing in the charter, imo, is that the spec
should specify what is the appropriate behaviour when an application is
not able to handle some rule or ruleset or when it is unable to
determine whether it can handle them correctly (meaning: as they were
intended to be). Indeed, fully implementing the standard does not have
to mean being able to use everything that can be written with it.
Received on Tuesday, 23 August 2005 17:54:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:48:33 UTC