W3C home > Mailing lists > Public > public-xg-urw3@w3.org > July 2007

RE: [URW3 public] Re: [URW3] ... three questions based on the last telecon

From: Giorgos Stoilos <gstoil@image.ece.ntua.gr>
Date: Mon, 16 Jul 2007 20:26:29 +0300
Message-Id: <200707161726.l6GHQPlZ031057@manolito.image.ece.ntua.gr>
To: <Peter.Vojtas@mff.cuni.cz>, <public-xg-urw3@w3.org>

> -----Original Message-----
> From: public-xg-urw3-request@w3.org [mailto:public-xg-urw3-request@w3.org]
> On Behalf Of Peter Vojtas
> Sent: Monday, July 16, 2007 3:51 PM
> To: public-xg-urw3@w3.org
> Subject: Re: [URW3 public] Re: [URW3] ... three questions based on the
> last telecon
> The public list seems to work,
> Dear all, first of all I am very happy about this discussion (especially
> before the URSW deadline (who is going to submit what?)) and thankfully
> for answers.
>       First of all, syntax is not important for me, I was speaking
> conceptually, what I refer to, is the binary model (like RDF, OWL or
> most of DL's (a poor first order relative)). So far we can represent the
> Thing by an oriented graph - it is there, even XML is an oriented graph
> (under some assumptions).
>       I am not suggesting to use reification, I just used it in an
> example. My point is simple, where we (XG-URW3) are going - to convince
> people that we need extension of W3C standards - how to achieve the goal

I think that people are more-or-less convinced about the need of uncertainty
in Semantics Web (after 2 very difficult years in Knowledge Web
http://knowledgeweb.semanticweb.org/, a tough W3C Rule Languages workshop
http://www.w3.org/2004/12/rules-ws/report/minutes#discussion0, continuation
in RIF Chartering http://www.w3.org/2005/rules/wg/charter.html#extensibility
- sorry for not having the email pointers in hand - and two tough OWL ED
Workshops - OWL ED 2005 & 2007) otherwise we would not be talking through
this mailing list :).

What we need to do now is to prove that uncertainty representation and
reasoning in the Web is something *practically feasible*, i.e. there are
tools that are working, uncertainty-ontologies. 

> - by use cases and/or examples - I have used reification to explain that
> in our model we need to have both the structure of
> (sentence/proposition/statement) and speaking about these - e.g. who is
> the author, who (person/tool) assigned a uncertainty degree to it,
> ....(when - our uncertainty about the statement can change in time..)
> and it is already possible using W3C standards.

OK...but then where is the new extension we need to convince for?

>       Of course syntax can and is important when handling syntactical
> object and some notations make it easier (like RDF/XML) or even the DOM
> model of it.
>       I think RDF is quite expressive (I do not say optimal, effective
> nor easy to read) because subject - predicate - object REFERS TO
> linguistic subject - verb - object, and so far humans are able (or at
> least some try to) understand each other, it is not so bad.
>      Of course I know from databases that every relation can be
> decomposed to several binary, but the price (to effectiveness) to pay is
> large. I have colleagues which developed a small generalization of RDF,
> a "heap data model" where instead of three columns they have some more
> for most common reifications - time stamp, author,... of the triple
> hidden in first three entries of the relation row.
>        So I agree to use some more readable notation than triples (but
> conforming with W3C standards).

Ok. Lets take OWL then and see your examples there.

>       By the way, even if "the Ontology is *not* meant to describe how
> to capture uncertainty in practice." then I ask: what is a measure of
> ontology being relevant to our problems. I think we already have
> reification in our model of uncertainty, namely in the the fact that the
> uncertainty is attached to a sentence (or as is our discussion some say
> rather to proposition). I argue that even the triple
>         urw3:Sentence      urw3:hasUncertainty  urw3:Uncertainty
> is a subject to further statements, e.g. who said this.

You always make the assumption that the basic structure we have is a triple
and everything is interpreted as such. In OWL you can probably use other
things (rather than reification of triples), like annotations to state these
FYI, we recently succeeded in providing an RDF serialization of fuzzy-OWL
without using reification. 


>       Our charter says "However, the entire use case collection would
> provide a basis for discussing whether the recommended set is sufficient
> to advocate further actions along the W3C Recommendation Track, either
> as a separate Recommendation or as part of other related work." I
> advocate for compatibility with W3C standards (maybe via some
> transformations, mappings, extensions,...at least in the sense, that a
> fully certain model is the W3C one).
>       To conclude: My point of view is, that the information on the web
> is not attached with an uncertainty as it is, it always depends on the
> interpreter, how much do I know about the author (security, trust
> issues), what is my background knowledge, ... in the social web - who
> else attached an uncertainty to this, how respectable, trustworthy these
> are,...
>       And last, I do not speak about uncertainty examples, which are
> uncertain even when not on the web - like diagnosis in medicine or the
> weather :-)
>       greetings Peter
Received on Monday, 16 July 2007 17:26:48 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:50:54 UTC