W3C home > Mailing lists > Public > www-tag@w3.org > July 2003

Re: resources and URIs

From: pat hayes <phayes@ihmc.us>
Date: Sat, 19 Jul 2003 15:19:24 -0500
Message-Id: <p06001237bb3f49de4ce2@[]>
To: Michael Mealling <michael@neonym.net>
Cc: www-tag@w3.org

>On Wed, 2003-07-16 at 13:17, pat hayes wrote:
>>  >I think it is reasonable to say that, if the Web architecture document
>>  >is to embody all of the SW and WSA activities, the Introduction
>>  >and most of the document will have to be rewritten.  It is also
>>  >reasonable to say that we haven't gotten that far yet, nor do we
>>  >intend to until after this version is released.
>>  Ok, fair enough: but could y'all at least try to *not* say things
>>  that will cause actual harm? For example, just avoiding the phrase
>>  "the resource of the URI" might be a great help, and instead using
>>  some phrase which is less likely to be understood as meaning 'the
>>  resource referred to by the URI'.
>Why does an existing, well deployed set of standards and systems have to
>change its lexicon to suit a system that has only mildly been specified
>and rarely deployed?

Because the new 'system'illuminates the fact that the terminology is, 
and always has been, completely confused?

If this were a genuine technical lexicon, then fine: but it is not. 
The W3C has *nowhere*, *ever* , published a coherent and meaningful 
explanation of what these words and assertions mean.  RFC 2396 does 
not do so, the TAG architecture document does not do so. The words 
"resource" and the claim that URIs "specify" them uniquely is just 
repeated almost in the exact same words again and again, never with 
any exposition or explanation, or even any indication of why it is 
felt so strongly that specification must be unique.  Sincere attempts 
to discover what they mean have produced only emotional denials that 
anything needs to explained, or that the meanings can be found in any 
dictionary.  The meanings suggested by the illustrative examples 
offered seem to point in different and conflicting directions: it is 
hard or impossible to see what they have in common. Suggested 
interpretations ( I have at various times suggested three, I believe) 
have always been rejected as mistaken or inadequate, but no-one seems 
able to offer an alternative.  These are not ordinary English words 
being used casually, but they are not defined or elaborated anywhere. 
To claim that this is a deployed lexicon is disingenuous.

Most of the things said using this lexicon seem to be false, in any 
case.  For example, existing Web practice does not depend on URIs 
having unique denotations: the very fact that current practitioners 
can disagree casually about whether or not it is OK to say that an 
email address denotes a mailbox or a person, or can (more often) not 
give a damn, seems to me to show clearly that the current 
architecture and design in fact do NOT depend on this semantic 
conditions. In the case of email addresses, for example, the job of 
the architecture is to deliver mail properly. What the URI is used to 
*indicate* is irrelevant to the (pre-SW) architecture, and probably 
impossible to specify without more knowledge of the context of use.

>  Don't get me wrong, I'm a big fan of the SW and I
>really hope to use it in lots of applications. But I've spent far to
>much time getting ordinary standards setting people up to speed on what
>we've already specified to change the language (and meaning) out from
>under them....

I actually strongly suspect that the meaning doesn't need  be 
changed, in a sense, but you aren't giving an actual specification 
when you use words that are not defined and leave them undefined. 
Certainly I don't want to alter the architecture; but I don't think 
that your description is accurate as it stands or does the Web 
justice. In fact, the Web uses URIs with much of the flexibility that 
denoting expressions are used in all human languages, and gets much 
of its power and utility from that. A language which insisted on 
single denotations would be a primitive language rather like archaic 
Chinese or ancient Egyptian heiroglyph languages, where as far as 
possible every concept was given a unique glyph. Fortunatly, language 
has evolved since then. Even RDF is more sophisticated than that.  It 
is OK for an email address to be systematically ambiguous; that is an 
efficient and proper use of language.  It is almost impossible to 
stop people using language in this way. The only thing that is wrong 
is having some central authority which emits authoritative 
pronouncements which if actually followed strictly would probably 
render the Web inoperative in a matter of months.

>IMHO, the best thing the SW web could do is create a brand new word and
>define it to mean whatever it is they want it to mean and leave the rest
>of the architecture as is....

The SW will use the same Web architecture as other Web applications 
use.  And my naggings have nothing directly to do with the semantic 
web.  All my examples were taken from the existing Web and the TAG 
architecture document itself.  My goal is to try to have the Web 
architecture described accurately without getting this description 
tangled up with weird, unnecessary and unworkable semantic 

Pat Hayes

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 Saturday, 19 July 2003 16:19:30 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:00 UTC