W3C home > Mailing lists > Public > public-swbp-wg@w3.org > March 2005

R:comment - RDFTM: Survey of Interoperability Proposals

From: Fabio Vitali <fabio@cs.unibo.it>
Date: Thu, 10 Mar 2005 03:04:31 +0100
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, <bernard.vatant@mondeca.com>
Cc: <public-swbp-wg@w3.org>
Message-ID: <BE5569BF.1178F%fabio@cs.unibo.it>

Hi. 

First of all, let me provide my sincerest excuses to the WG for not
providing my opinions earlier and for not taking part to the discussion in
the last few days. I know that my presence (or, rather, lack thereof) in the
WG mailing list has been noted and expressed, and for this I am also deeply
sorry. 

Unfortunately, I have no strong opinion on the fundamental basis of RDF and
TM (they are tools to an end for me) and thus am somewhat overawed when
deep, technical, competent arguments are brought forth. Since I do not have
strong opinions, I need time to form one when competent, technical arguments
are brought forth, and even more time to write it down. So please bear with
me, and do not expect continuous and impetuous input from me. I just hope
that, whenever I manage to get something in type, it might be found
interesting for the group, unless in the meantime someone else has
progressed the discussion away from my contribution, a fact which, given the
speed of my thoughts, seems in fact quite likely. :-)

So, while considering some of the comments that our draft has brought forth,
let me just specify one small facts:

In the charter (can I call it a charter?) of the RDFTM task force
(http://www.w3.org/2001/sw/BestPractices/RDFTM/), these are listed as our
deliverables: 
* A Working Group Note describing the pros and cons of existing approaches.
* A Working Group Note or Recommendation describing guidelines for
transforming a topic map into an RDF/OWL representation and vice versa.

The charter (I will call it a charter, then) assumes without discussions
that interoperability of RDF and TM is in fact relevant to the activity of
the full SWBPD working group. There is no mention of the relative merits of
the two approaches, and no discussion of the possible pointlessness of the
enterprise. 

That it is interesting to interoperate or convert from one language to the
other is given as a fact, and in my opinion this is true without discussion,
given the fact, as Bernard points out, that there *are* people out there
using TM for expressing metadata about Web resources, and at some point it
might be interesting to draw whatever inference we might, even if
inconclusive or potentially non terminating, from a collection of TM topics,
and to merge them with some other (formally superiorly defined) collection
of RDF statements. 

In fact, that interoperability or conversion might be difficult or
impossible to reach given the current poor status of the underlying
formalism of TM is to be considered a result (or, possibly, a non-result) of
the activity of the TF, and not one of its starting tenets, in my opinion.

The second consideration I want to share with you is the fact that the first
output of the TF is a document describing the pros and cons of the existing
approaches. This means a SURVEY of *existing*, *external*, *unendorsed*
proposals (er, yes, the two proposals that come out best of the document are
in fact the ones that were proposed by the authors of the survey themselves,
but I hope you agree that this does imply endorsement in any form).

Since it is a survey, it is not its purpose to describe anything more than
what is needed to be known in order to understand the existing proposals in
literature. As a survey, and not a position statement, it does not deal with
RDFS, OWL, or anything else, inasmuch the surveyed works deal with them
(which they do not).

So this leads me to conclude that while I believe that discussion about
RDFTM interoperability is a fruitful activity to do, and something we might
want to spend some time upon, this should not be done while providing
feedback and input for the interoperability survey, which is an activity on
a different level altogether, and more or less independent on the relative
merits or feasibility of the interoperability model itself.

And secondly, I tend to believe that the identification of the right
vocabulary and technical panoply to provide an actual interoperability
guideline is matter of the second activity of the TF, not this one. Whether
RDFS, OWL, or anything else will be relevant for the actual conversion
should be matter of discussion from now on.

> This lack of a definition matters for reasons from both the RDF and the
> Topic Maps side.  RDF has undergone a significant change in the last few
> years from a pre-theoretic language with no firm foundation (see Resource
> Description Framework (RDF) Model and Syntax Specification
> http://www.w3.org/TR/1999/REC-rdf-syntax-19990222/) to a full-fledged logic
> (see RDF Semantics http://www.w3.org/TR/rdf-mt/).  Which version of RDF is
> meant in the document?  Which version of RDF to the interoperability
> proposals refer to?

The only relevant information in this context is which version of RDF and TM
do the surveyed work refer to. Whether they are the most appropriate or
complete references are not relevant in this context (except, possibly, to
point out that they refer to older and inappropriate versions of the
relevant standards).

If the group agree, we will add references to the versions of the documents
each proposal refer to. I personally doubt that it might make any difference
or shed any light on the issues discussed. But if this is the opinion of the
group, I will happily concede.


>As well, what is the place of RDFS in the document?
> Is it included?  Is it excluded?

It is not relevant inasmuch none of the surveyed proposals mention it as
relevant for the translation.

> Which
> version of Topic Maps is under consideration?  Does it matter?

Whichever version has been considered in each proposals. And no, I don't
think it matters. But if we think it useful, we will examine and try to
deduce which specific version of TM each proposal was referring to.

> The second problem is that many of the interoperability proposals predate
> the finalization of the RDF Semantics.  Their current applicability is thus
> very suspect.  The document needs to carefully consider this aspect of each
> proposal. 

I am not sure of what this means. It looks like an objection on principle
rather than a deduction on the actual results of the survey. Surely nothing
in the discussion of the proposals has any relevance to the current status
of the RDF language. Possibly the final part, where we provide some
recommendations, might, but this can be deduced quite clearly by reading its
content. Is anything on the discussion we provide in chapter 4: Analysis,
that is made incorrect or superfluous by the finalization of the RDF
semantics?

>>> The third problem is that RDF and Topic Maps belong to different
>>> categories, at least so far as I can determine.
>> 
>> By essence, yes. But their story tends to show that differences in nature
>> seem less and
>> less important compared to the community of what they both want to achieve.
>> People do
>> things with TM that they could do with RDF, and vice-versa. Maybe they should
>> not, maybe
>> it's hacking, in any case they do it. Why? Only out of poor understanding and
>> mental
>> confusion, like football fanatics?
> 
> So what?  If RDF and Topic Maps belong to different categories and are thus
> incomparable, then there are two ways to proceed.  The first way is to
> simply realize this, and not attempt to achieve (or document)
> interoperability.  The second way is to augment or modify one or the other
> so that interoperability is possible.
> 
>>> RDF is now a formally-specified logic with a model-theoretic semantics.
>>> Topic Maps is not.  This difference matters, and needs to be taken into
>>> account in every
>>> discussion of the relationship between RDF and Topic Maps.
>> 
>> Agreed again.
>> 
>>> At best, there needs to be some way to determine that the interoperability
>>> proposals
>>> preserve logical equivalence on the RDF side.  At worst, there is no point
>>> in doing any mappings, as RDF and Topic Maps are simply incomparable.
>> 
>> Sorry but I think this latter position is not more sustainable that saying,
>> say, English
>> and Chinese are simply incomparable.   Sure there are many things in Chinese
>> that are not
>> translatable in English, and maybe the other way round, but that does not
>> mean there is no
>> point in trying to make Chinese-speaking and English-speaking folks trying to
>> understand
>> each other.
> 
> Ah, here is where I disagree.  English and Chinese fit into the same
> category, namely human natural languages.  There is at least therefore some
> evidence that they are comparable.  I don't see any such evidence for RDF
> and Topic Maps.

I am not sure I understand. The important bit about languages is to find out
whether people use them for the same purposes, not that they have been built
using the same principles or with the same level of formalism. It is a
matter of usage, not grammar. Is TM used for very similar purposes as RDF? I
think so. Therefore it is relevant to provide translation. Which will be
limited and inadequate and incomplete because of the difference in
formalism, but it will be better than no translation at all, if we really
believe that interoperability is a goal.

Finally, I would propose that further discussions should be clearly marked
as belonging to one and only one of the following three categories:

* The existence of RDFTM itself: whether there is any point in proceeding
with the interoperability guidelines given the complex differences between
RDF and TM approaches
* Modifications to the survey: What are the limits of the survey considered
as a survey, and not as a proposal, that need to be considered and solved
for the document to reach a final status
* Preparation for the guidelines: Supposing that we in fact decide to go on
with the RDFTM TF, what additional languages, theories, models and
approaches need to be considered in the next step of the activities of the
TF, namely in the delivery of an actual set of guidelines for
interoperability, that weren't discussed in the proposals surveyed.

Well, I guess that's all I had to say in the meantime.

Ciao a tutti e buonanotte (3:04 AM).

Fabio Vitali

--

Fabio Vitali                            Tiger got to hunt, bird got to fly,
Dept. of Computer Science        Man got to sit and wonder "Why, why, why?'
Univ. of Bologna  ITALY               Tiger got to sleep, bird got to land,
phone:  +39 051 2094872              Man got to tell himself he understand.
e-mail: fabio@cs.unibo.it                     Kurt Vonnegut, "Cat's cradle"
http://www.cs.unibo.it/people/faculty/fabio/
Received on Thursday, 10 March 2005 02:05:04 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:31:07 UTC