Moving this dialog to the [ontolog-forum] list here (with Enrico and
Pat's consent), so it becomes part of the ongoing open dialog, and
gets captured for the benefit of the entire community and future
reference. =ppy (01)
---------- original message ----------
From: Pat Hayes <phayes@xxxxxxx>
Date: Sat, Nov 23, 2013 at 11:58 AM
Subject: Re: Proceedings: "Rules-Reasoning-LP" mini-series session-03
- Thu 2013.11.21
To: Enrico Franconi <franconi@xxxxxxxxxxxx>
Cc: Peter Yim <peter.yim@xxxxxxxx> (02)
On Nov 23, 2013, at 3:40 AM, Enrico Franconi <franconi@xxxxxxxxxxxx> wrote: (03)
> Dear Pat,
> I believe that there is one expected obvious meaning to the notion of
>"subClass", which corresponds to subset of the corresponding extensions if you
>see classes as sets. (04)
Yes, but RDFS (quite explicitly) does NOT see classes as sets. Of
course, considering classes to be sets gives you the extensional
semantics, but that is not how the RDFS data model was designed. (05)
> As a matter of facts, this is partially true in the RDFS semantics, namely
>that rdfs:subClass implies the subset of the corresponding extensions, but not
>viceversa. (06)
Yes, exactly. (07)
> But I'm not willing to argue about what is obvious and what is not, since
>there may be different categories of 'obvious', so let's not go on with this
>potentially infinite discussion :-) (08)
We are not having a discussion. I am telling you why the RDF(S) data
model was designed in the way that it was, which involves not treating
classes as identical to their set extensions, and not using your 'iff'
definitions of subProperty and subClass. This was not an inadvertant
omission, but a deliberate design decision. I believe that the 2004
Semantics document makes this quite clear. This has now become a
non-issue, and all mention of extensionality in RDF(S) has been
dropped from the revised RDF 1.1 specifications, since as far as the
WG could determine, nobody has expressed the slightest interest in the
issue. (09)
> But let's focus on the inferences which are lost if you adopt a weak
>semantics as in RDFS. (010)
"Losing" inferences is actually an advantage, if you are not
particularly interested in the consquences of those inferences, as it
simplifies reasoning and saves time. The overwhelming majority of RDF
applications, and those for which it was primarily designed, involve
using subclass and subproperty and domain assertions to conclude
data-level facts from other data, not to derive new theorems at the
vocabulary-ontology level. If you want to do that, by all means use
OWL. But most RDF users have not even read the OWL specs. (011)
> The missed inferences I'm showing in slide 11 are actually expected by
>everyone I met so far. (012)
Then you have not met a wide enough class of users :-) I will concede
that the inference patterns (013)
aaa subProperty bbb .
bbb domain//range ccc .
|=
aaa domain//range ccc . (014)
are a reasonably intuitive idea which could be added as axioms,
although I can see ways in which it would be found puzzling by some
users (those who are used to thinking of domains and ranges as
attached to a class rather than to its extension.) But I have never
come across any application that needed it, and AFAIK it has never
been requested or even suggested by any RDF user. Part of the overall
design philosophy of RDF was to keep the 'core' semantics as minimal
as possible, and allow semantic extensions to extend this minimal
core. (015)
> Similarly, there are analogous missed inferences with 'domain'. In all my
>discussions with people around the world, I always had a surprise reaction;
>people never realised these inferences were missing and they wonder why they
>don't have them. People build applications by assuming the presence of all
>valid statements, but here they are missing some. (016)
No, they are not missing valid statements, because these are not valid
with the semantics as stated. Calling them "valid" begs the question.
I am sure many people, especially people with a logical training, do
indeed come to RDFS with a presumption that the semantics is
extensional, and are surprised to find that it is not. However, many
other people (for example, users trained in library science) find such
conclusions quite unintuitive, and find other consequences of
extensionality, such as the identity of classes with the same members,
not only unintuitive but clearly wrong. (017)
> So, I don't really understand your statement that extensional RDFS "has
>entailments that are not at all natural in many cases, and would make problems
>for a number of widely used ontologies". Can you please elaborate on that? (018)
The data model of SKOS, for example, is explicitly non-extensional.
But also, identifying classes and properties with their extensions
would make many RDF axioms incoherent, for example (019)
rdfs:Class rdf:type rdfs:Class . (020)
would violate the ZFC axiom of foundation. (021)
Pat (022)
> cheers
> —e. (023)
======== (024)
---------- original message ----------
From: Enrico Franconi <franconi@xxxxxxxxxxxx>
Date: Sat, Nov 23, 2013 at 1:45 AM
Subject: Re: Proceedings: "Rules-Reasoning-LP" mini-series session-03
- Thu 2013.11.21
To: Pat Hayes <phayes@xxxxxxx>
Cc: Peter Yim <peter.yim@xxxxxxxx> (025)
I forgot to add a note regarding your comment on slide 8. That slide
has the only purpose to show the necessity of explicit transitivity
axioms. So, you are right that the "conclusion follows from the
transitivity of rdfs:subProperty", since this is exactly what I wanted
to show. (026)
cheers
—e. (027)
> On 23 Nov 2013, at 10:40, Enrico Franconi <franconi@xxxxxxxxxxxx> wrote: (028)
Dear Pat, (029)
I believe that there is one expected obvious meaning to the notion of
"subClass", which corresponds to subset of the corresponding
extensions if you see classes as sets. As a matter of facts, this is
partially true in the RDFS semantics, namely that rdfs:subClass
implies the subset of the corresponding extensions, but not viceversa. (030)
But I'm not willing to argue about what is obvious and what is not,
since there may be different categories of 'obvious', so let's not go
on with this potentially infinite discussion :-) (031)
But let's focus on the inferences which are lost if you adopt a weak
semantics as in RDFS. The missed inferences I'm showing in slide 11
are actually expected by everyone I met so far. Similarly, there are
analogous missed inferences with 'domain'. In all my discussions with
people around the world, I always had a surprise reaction; people
never realised these inferences were missing and they wonder why they
don't have them. People build applications by assuming the presence of
all valid statements, but here they are missing some. (032)
So, I don't really understand your statement that extensional RDFS
"has entailments that are not at all natural in many cases, and would
make problems for a number of widely used ontologies". Can you please
elaborate on that? (033)
cheers
—e. (034)
============ (035)
---------- original message ----------
From: Pat Hayes <phayes@xxxxxxx>
Date: Sat, Nov 23, 2013 at 12:14 AM
Subject: Re: Proceedings: "Rules-Reasoning-LP" mini-series session-03
- Thu 2013.11.21
To: Peter Yim <peter.yim@xxxxxxxx>
Cc: Enrico Franconi <franconi@xxxxxxxxxxxx> (036)
On Nov 22, 2013, at 11:04 PM, Peter Yim <peter.yim@xxxxxxxx> wrote: (037)
> Dear Pat,
>
>
> During yesterday's RulesReasoningLP mini-series session-03, Enrico
> Franconi made a very interesting presentation, entitled: "The Logic of
> Extensional RDFS" during which he remarked on certain aspects of
> RDF/RDFS for which he may have an elegant solution for. (He even
> called out your name, Pat, in the talk.)
>
> See: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_11_21#nid412U
> ... with his slides at:
>
>http://ontolog.cim3.net/file/work/RulesReasoningLP/2013-11-21_Concepts-Foundations-II/The-Logic-of-Extensional-RDFS--EnricoFranconi_20131121.pdf
> ... and recording of his talk, starting at time-point 1:03:56 in
>
>http://ontolog.cim3.net/file/work/RulesReasoningLP/2013-11-21_Concepts-Foundations-II/RulesReasoningLP-s03_20131121b.mp3
> ...... ref. also my exchange with him (asking if you are aware of this
> work) at time-point 1:22:20
>
> I promised I will write you, to help draw your attention to what he
> has been working on, and maybe you can help assess whether it would be
> meaningful to incorporate some of his thinking into the W3C RDF/RDFS
> standardization exercise. I do understand that the timing is probably
> off, and any candidate change has probably missed the cycle and
> iteration that you've been working on over the last couple of years
> (or so) ... (038)
Yes, the timing is quite impossible. These documents are now past the
last call stage and are in candidate recommendation. AFAIK, Enrico has
not made any of this visible to the RDF WG in any comments. So this is
basically irrelevant until the W3C calls for a new RDF WG, which will
likely not happen now for several years. (039)
FWIW, we did consider an extensional semantics for RDF in 2004 but
rejected it as it seemed less natural than the intensional semantics.
It has entailments that are not at all natural in many cases, and
would make problems for a number of widely used ontologies (notably
SKOS). I argued, unsuccessfully, for OWL to adopt the intensional
semantic style, which I still think is more suitable for most data
applications. The 2004 RDF semantics did outline this extensional
semantics (non-normatively) and explained some of its consequences.
The level of interest in this in the RDF world has been so low that we
simply omitted it from the 2013 specifications, but it is of course a
valid semantic extension of RDFS. (040)
I have not yet had the chance to listen to the talk, but glancing at
the slides, slide 5 is incorrect. This is not the "intended meaning"
of rdfs:subClass. If it were, of course the published semantics would
be inadequate. The rest of the presentation seems to be solving a
problem which is not actually a problem, under the misapprehension
that RDFS was intended to be an extensional logic in the first place.
But it was not. (041)
Slide 8 puzzles me. The conclusion would seem to follow from the
transitivity of rdfs:subProperty, unless the red diamond has more
meaning than is apparent in the slide (?) (042)
I will comment further when I have had a chance to listen to the talk. (043)
Pat (044)
>
> Comments, pointers, ... ?
>
>
> Regards. =ppy
> -- (045)
------------------------------------------------------------
IHMC (850)434 8903 home
40 South Alcaniz St. (850)202 4416 office
Pensacola (850)202 4440 fax
FL 32502 (850)291 0667 mobile (preferred)
phayes@xxxxxxx http://www.ihmc.us/users/phayes (046)
========== (047)
---------- Forwarded message ----------
From: Peter Yim <peter.yim@xxxxxxxx>
Date: Fri, Nov 22, 2013 at 9:04 PM
Subject: Fwd: Proceedings: "Rules-Reasoning-LP" mini-series session-03
- Thu 2013.11.21
To: Pat Hayes <phayes@xxxxxxx>
Cc: Enrico Franconi <franconi@xxxxxxxxxxxx>, Peter Yim <peter.yim@xxxxxxxx> (048)
Dear Pat, (049)
During yesterday's RulesReasoningLP mini-series session-03, Enrico
Franconi made a very interesting presentation, entitled: "The Logic of
Extensional RDFS" during which he remarked on certain aspects of
RDF/RDFS for which he may have an elegant solution for. (He even
called out your name, Pat, in the talk.) (050)
See: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_11_21#nid412U
... with his slides at:
http://ontolog.cim3.net/file/work/RulesReasoningLP/2013-11-21_Concepts-Foundations-II/The-Logic-of-Extensional-RDFS--EnricoFranconi_20131121.pdf
... and recording of his talk, starting at time-point 1:03:56 in
http://ontolog.cim3.net/file/work/RulesReasoningLP/2013-11-21_Concepts-Foundations-II/RulesReasoningLP-s03_20131121b.mp3
...... ref. also my exchange with him (asking if you are aware of this
work) at time-point 1:22:20 (051)
I promised I will write you, to help draw your attention to what he
has been working on, and maybe you can help assess whether it would be
meaningful to incorporate some of his thinking into the W3C RDF/RDFS
standardization exercise. I do understand that the timing is probably
off, and any candidate change has probably missed the cycle and
iteration that you've been working on over the last couple of years
(or so) ... (052)
Comments, pointers, ... ? (053)
Regards. =ppy
-- (054)
---------- Forwarded message ----------
From: Peter Yim <peter.yim@xxxxxxxx>
Date: Fri, Nov 22, 2013 at 8:48 PM
Subject: Proceedings: "Rules-Reasoning-LP" mini-series session-03 -
Thu 2013.11.21
To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>, Markus
Kroetzsch <markus.kroetzsch@xxxxxxxxxxxxx>, Hector Perez Urbina
<hector@xxxxxxxxxxxxxxx>, Hassan Ait-Kaci <hassanaitkaci@xxxxxxxxx>,
Enrico Franconi <franconi@xxxxxxxxxxxx>
Cc: "Obrst, Leo J." <lobrst@xxxxxxxxx>, Pascal Hitzler
<pascal.hitzler@xxxxxxxxxx>, Benjamin Grosof
<benjamin.grosof@xxxxxxxxxxxxxxxxxxxxx>, Evren Sirin
<evren@xxxxxxxxxxxxxxx>, Peter Yim <peter.yim@xxxxxxxx> (055)
Dear All, (056)
We had a really interesting and substantive session yesterday. I'm
writing to let everyone know that the full proceedings of the session
- agenda, abstracts, slides, in-session chat-transcript, audio
recording, attendee roster, etc. - are all on archive now, available
online, and publicly accessible under:
http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_11_21#nid3ZV5 (057)
A huge THANK YOU to all the speakers for sharing their work and
insights with us. They are, again, contributing key pieces to the body
of knowledge that this community has been collectively building. (058)
The next event (session-04) for this RulesReasoningLP mini-series will
be coming up on Dec-19, entitled: "Guide to Reasoning Applications
Development and Cases" - Co-chairs: HensonGraves & KenBaclawski -
watch out for details in further announcements, and posts on the wiki
under "News and Announcements" -
http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nidW (059)
In the mean time, Happy Thanksgiving (to those in the US.) There will
not be any mini-series panel session next week. Do join us, though,
in two weeks (Thu 2013-12-05) for the OntologySummit2014 Pre-Launch
Community Session, when we will collaboratively work up a program for
our next OntologySummit.Talk to you all, then! (060)
Regards. =ppy (061)
For the Session Co-chairs
Leo Obrst & Pascal Hitzler (062)
... along with other co-champions of
the RulesReasoningLP mini-series
-- (063)
---------- Forwarded message ----------
From: Peter Yim <peter.yim@xxxxxxxx>
Date: Wed, Nov 20, 2013 at 9:36 AM
Subject: Re: "Rules-Reasoning-LP" mini-series session-03 - Thu 2013.11.21
To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx> (064)
.
http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_11_21 (065)
= REMINDER = (066)
This is a reminder that the "RulesReasoningLP" mini-series session-03:
"Concepts and Foundations of Rules and Ontologies: Logic Programs,
Classical Logic, and Semantic Web - II " is coming up tomorrow
(Thursday, Nov-21.) (067)
Dr. Leo Obrst and Professor Pascal Hitzler will be co-chairing the
session, and have invited over Dr. Markus Kroetzsch, Dr. Hector
Perez-Urbina, Professor Hassan Ait-Kaci and Professor Enrico Franconi
to join us as panelists. An open Q&A and discussion will follow, as
usual, after the panelists' presentations. (068)
If you have a professional interest in Ontology, Rules, and Logic
Programming for Reasoning and Applications, you probably don't want to
miss this session, and the rest of the events of this mini-series. (069)
RSVP now, if you haven't already! (... see details below) (070)
* Date: Thursday, 21-Nov-2013 (071)
* Start Time: 9:30am PST / 12:30pm EST / 6:30pm CET / 17:30 GMT/UTC
ref: World Clock -
http://www.timeanddate.com/worldclock/fixedtime.html?month=11&day=21&year=2013&hour=9&min=30&sec=0&p1=224
* Expected Call Duration: ~2.0 hours (072)
* Dial-in:
** Phone (US): +1 (206) 402-0100 ... Conference ID: 141184# , or
** Skype: in view of recent reported skype connection issues, this is
not recommended (especially for speakers) although it may still
work for some ... to do that, make a skype call to
the skypeID: "joinconference"
... when prompted enter Conference ID: 141184# (073)
Just point your browser to the session page:
http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_11_21 when
you are about to connect into the session. Details will be there at
the top of the page. Please try dialing-in a few minutes before the
scheduled start-time, as we have a very full program planned for this session. (074)
Talk to you all tomorrow! =ppy (075)
For the Session Co-chairs
Leo Obrst & Pascal Hitzler (076)
... along with other co-champions of
the RulesReasoningLP mini-series
-- (077)
---------- original message ----------
From: Peter Yim <peter.yim@xxxxxxxx>
Date: Fri, Nov 15, 2013 at 3:27 PM
Subject: "Rules-Reasoning-LP" mini-series session-03 - Thu 2013.11.21
To: "[ontolog-invitation]" <ontolog-invitation@xxxxxxxxxxxxxxxx> (078)
.
Re: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_11_21 (079)
Dear All, (080)
After the two very successful sessions last month, we are pleased to
feature virtual panel session-03 of the "Ontology, Rules, and Logic
Programming for Reasoning and Applications" (dubbed:
"RulesReasoningLP") mini-series next Thursday, 21-Nov-2013. You are
cordially invited to join us ... RSVP Now (details below)! (081)
* Date: Thursday, 21-Nov-2013
* Start Time: 9:30am PST / 12:30pm EST / 6:30pm CET / 17:30 GMT/UTC
** ref: World Clock -
http://www.timeanddate.com/worldclock/fixedtime.html?month=11&day=31&year=2013&hour=9&min=30&sec=0&p1=224
* Expected Call Duration: ~2.0 hours (082)
*Register your attendance* by emailing <peter.yim@xxxxxxxx> off-line
or register yourself directly to the wiki session page. Please specify
the date or name of the session(s) you are registering for (your
affiliation too, if you are not already a member of the community.) (083)
= RulesReasoningLP: mini-series session-03 - Thu 2013-11-21 = (084)
Topic: Concepts and Foundations of Rules and Ontologies: Logic
Programs, Classical Logic, and Semantic Web - II (085)
Session Co-chairs: Dr. Leo Obrst (Ontolog; MITRE) & Professor Pascal
Hitzler (Wright State U) (086)
Panelists / Briefings: (087)
* Dr. Markus Kroetzsch (Technische Universitat Dresden) - "Existential
Rules in Ontological Modelling" (088)
* Dr. Hector Perez-Urbina (Clark & Parsia, LLC) - "Modeling with Rules
in Practice" (089)
* Professor Hassan Ait-Kaci (Universite Claude Bernard Lyon 1) -
"Reasoning and the Semantic Web" (090)
* Professor Enrico Franconi (Free University of Bozen-Bolzano) - "The
Logic of Extensional RDFS" (091)
This session is co-chaired by Dr. Leo Obrst and Professor Pascal
Hitzler, and will be the second of the two sessions devoted to
addressing the concepts and foundations of the technologies underlying
ontology and rule reasoning, especially focused on logic programming
and Semantic Web extensions. The session will comprise of briefings
from the four invited panelists, followed by an open discussion among
the panel and all the participants. See details on the session page
at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_11_21 (092)
This session, like all Ontolog virtual events, are open and free of
charge. Anyone who is interested, or (better still) who may have
something to contribute, is welcome. Please refer to event details on
the session pages, to which the hyperlink is given above, where you
will find session agenda, conference call dial-in, slides and other
pertinent information. Feel free to pass this invitation along to
colleagues who may also find these sessions to be of interest. (093)
*RSVP* by emailing Peter Yim at <peter.yim@xxxxxxxx> offline (or add
yourself directly to the session page if you are already an Ontolog
community member) so that we can prepare enough resources to support
everyone. [Please state clearly the date of the session you are
registering for in your email; your affiliation too, if you are not a
community member.] (094)
These sessions will be recorded and made available in a publicly
accessible archive. Therefore, before participating, please make sure
you are cognizant of our IPR policy (ref:
http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid32). (095)
Regards. =ppy (096)
For the Session Co-chairs
Leo Obrst & Pascal Hitzler (097)
... along with other co-champions of
the RulesReasoningLP mini-series (098)
p.s. see more details on the mini-series at:
http://ontolog.cim3.net/cgi-bin/wiki.pl?RulesReasoningLP
where you will find links to the full proceedings (including slides,
in-session-chat-transcript,
and audio recording) of session-01 & -02 at their respective session pages.
-- (099)
_________________________________________________________________
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 (0100)
|