Re: update on I18N review of RDF, OWL

>At 09:41 03/06/26 -0500, Dan Connolly wrote:
>
>>On Mon, 2003-06-16 at 17:39, Martin Duerst wrote:
>>>  Hello Dan,
>>>
>>>  Many thanks for contacting me on this.
>>>  Please see below for my take on your assumptions.
>>>  I have copied the I18N IG list, used for technical
>>>  discussions.
>>>
>>>  At 13:47 03/06/11 -0500, Dan Connolly wrote:
>>>  >Further to my action to get confirmation from I18N WG
>>>  >that our last call spec is I18N-happy...
>>>  >
>>>  >I talked with Martin in Budapest a couple weeks ago.
>>>  >
>>>  >Since then, he collected his thoughts on RDF literals
>>>  >and such...
>>>  >
>>>  >Summary of strings, markup, and language tagging in RDF (resend) Martin
>>>  >Duerst (Thu, Jun 05 2003)
>>>  >http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jun/0023.html
>>>  >
>>>  >and the RDF Core WG is working thru them.
>
>That didn't show very much, in particular it did not
>give any indication that the WebOnt WG thought about the
>possibility of equivalence ignoring language.
>
>One interesting discovery, though:
>
>http://lists.w3.org/Archives/Public/www-webont-wg/2002Nov/att-0223/01-dt.html
>
>It has been (deliberately?

Yes, deliberately.

>) left unclear as to whether a plain literal
>without a language tag is or is not an xsd:string.
>
>It would be good if this would be resolved the right way
>(namely that a plain literal without a language tag is
>an xsd:string). If that's not done in RDF, then at least
>in OWL (see below).

Why would it be good to resolve this in the 'right' way? It seems to 
me that there is room for genuine disagreement about this question, 
and given that there is such room, the best policy for a semantic 
language to take it to be agnostic about it. This isn't likely to 
cause any real problems; and if it does, they can be handled by 
special assumptions, possibly embedded in handling code, which can 
then be considered to be a semantic extension and therefore will be 
covered by the general semantic extendability conditions; whereas if 
we build this assumption into the semantics, then anyone  who wishes 
to reason within a strongly typed framework which distinguishes 
entities by their declared type (and there are such folk, and 
software built on such principles, eg Specware 
http://www.kestrel.edu/HTML/prototypes/specware.html) will be unable 
to proceed and probably unable to use the semantic web formalisms at 
all without awkward work-arounds. Taking a 'stance' on typing issues 
unnecessarily seems to be at odds with the general principles of 
semantic interoperability. Even the XML Schema Part 2 document is 
written using a model of typing which is very reminiscent of the 
strong typing model, eg it insists that value spaces of distinct 
types are disjoint, even when they are apparently defined to have the 
same members.

Pat Hayes

PS> The 'strong typing' view insists that entities - all entities - 
have an intrinsic 'type' which is part of how they are categorized, 
and that the 'same' thing is regarded as different when it is 
assigned a different type. It is widely used in computer science and 
underlies such ideas as distinguishing the integer zero from the real 
number zero from the complex number zero; it has respectable 
mathematical roots in topos theory and computational roots in 
programming language design, particularly in OO languages. On this 
view, to identify the bare character string 'abc' with the "same" 
string typed with a particular type, such as xsd:string, would be a 
category error like identifying the integer zero with the real number 
zero. I do not espouse this view, myself, but I recognize that it has 
been widely adopted and is in fact regarded as so obvious as to be 
hardly worth discussing by large sections of the technical and 
mathematical communities.
-- 
---------------------------------------------------------------------
IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes

Received on Wednesday, 2 July 2003 17:04:40 UTC