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

Re: resources and URIs

From: pat hayes <phayes@ihmc.us>
Date: Wed, 16 Jul 2003 14:46:21 -0500
Message-Id: <p06001221bb3b4fb7abdc@[]>
To: Tim Bray <tbray@textuality.com>
Cc: Jonathan Borden <jonathan@openhealth.org>, pat hayes <phayes@ihmc.us>, www-tag@w3.org

>Jonathan Borden wrote:
>>A few questions:
>>1) You are making a clear distinction between the "Web" and the "semantic
>>web". Is the TAG limited to "Web" issues, and is it the intention to define
>>a "Web" as distinct from a "SW" ... I ask, because as a member of the WebOnt
>>WG, I had assumed that the TAG was supposed to deal with issues involving
>>potentially *any* of the W3C WGs/activities. Is it your specific intention
>>to define an architecture for the "Web" as distinct from the "SW"? If so,
>>I'll stop commenting on any of this.
>One important differentiating characteristic is that today's Web 
>concretely exists, is rather well-debugged, and is daily use by 
>hundreds of millions.  I believe one important part of the TAG's 
>mandate is to write down the architectural principles that underly 
>this success story, which is all that the webarch doc has really 
>tried to do up to this point in time.

I absolutely agree; please don't understand my request for 
clarification as saying in any way that the Web is somehow faulty or 
that it needs to be replaced by a better one in order to get the SW 
working. My question/niggle way only about the *words* being used in 
the document, which *seemed* to be saying something that didn't 
*seem* to make sense.

>I am not against the TAG trying to contribute to documenting 
>architectural principles of the SW work (although I'd personally 
>have little to contribute); but that is going to be much harder than 
>documenting the empirically-observable principles that keep today's 
>web humming.

Indeed; but now, think about the weather-in-Oaxala example, which is 
entirely in the context of this current humming web. The document 
says that there is a "resource" which is (1) part of the Web (2) when 
pinged, delivers a representation, which is (3) "of" that resource. 

What makes the Web so useful in cases like this is indeed that the 
URI serves to access a representation (1 and 2); but it makes sense 
to ask of any representation  what it represents, what it refers to, 
what it is 'about' (in some sense) - and it seems obvious that if 
that is what is meant by the use of the words "representation of", 
then (3) is just plain false: and indeed that the Web *would not 
work* in the way that it does in fact work if it were true, since the 
only thing it would be good for would be to find out information 
about parts of the Web itself, not for example the weather in Oaxala.

So after reading things like this, being brought up short by finding 
what seems like obvious nonsense in a centrally important draft W3C 
technical document, and pondering what could possibly be the intended 
meaning of the words I must be somehow mis-reading, the best I can do 
is to surmise that what y'all mean by "representation of" is not at 
all what I (and many of your readers) would normally  understand this 
phrase to mean. You are not saying that the html-produced image on my 
screen is in any intuitive semantic sense 'about' the resource which 
is indicated by the URI. In my Chandra example, the stuff on my 
screen is a "representation of"  something on the WWW information 
network, not a "representation of" a galaxy far, far away.  Is that 
right? Because if so, then we do not need to say that a galaxy is a 
"resource", and everything is much simpler, and your document makes 
overall sense, although could perhaps have been better phrased.

>>If we are going to concern the "SW" at least as it is incarnated in current
>>activities and specific software products (i.e. RDF and OWL related
>>software), then you may certainly be a part of the Web without a URI. Hence
>>the statement is not correct in the SW context -- I'm not entirely convinced
>>that it is correct in the 'current Web' context e.g.
>This might be a nice clear clean differentiating principle, because 
>I'm pretty convinced that at the moment, something that doesn't have 
>a URI isn't part of the Web.

OK, suppose it isn't part of the Web. Can it still be *referred to* 
by something that is part of the Web? If so, then being 'part of the 
Web' is largely irrelevant to semantics, and we can agree about this 
and move on.  On this understanding, being "part of" the Web and 
being "referred to", or "described by", something on the Web are two 
completely different ideas.

>>Am *I* a part of the Web? I assert that I am a resource which is providing a
>>representation via a proxy that is part of the Web (by your definition).
>>That is to say my email address mailto:jonathan@openhealth.org is a URI by
>>which folks may communicate with *me* but I assert that it is not *my URI*
>>rather the URI of my mailbox. Is that a problem?
>That's not a problem problem.

But it IS a problem problem if you are going to talk semantics. Your 
document claims to be about architecture, but it is rife with 
semantic assertions (or at least it *seems* to be) and it uses 
language which is technical terminology in semantics. If you don't 
mean to be saying these things, then please try to use terms which 
don't seem to be making the claims that you seem to be making.

Maybe it would be worth remarking that semantics has been a fairly 
precise field, often using moderately advanced mathematical 
techniques, for about 70 years now. It isn't just De Saussure and 
Korzybsky any more.

>There is a resource identified by that URI.

That reads like a simple sentence and I am sure that you mean 
something simple and kind of obvious by saying it. However, honestly 
now, I have NO IDEA what you are saying. You need to spell it out 
better.  What is a 'resource' and what does it mean for a URI to 
'identify' it?

>  Today's Web doesn't have any way to talk formally and 
>machine-readably about what that resource is

Well, actually it does.  But in any case....

>, so the issue of whether it identifies Dr. Borden or his emailbox 
>is just not on the table.  Clearly the question of what the resource 
>is is of concern to humans

... it is of concern to *semantics*. And...

>and is readily dealt with in human language

.... no, it is NOT dealt with readily in normal language. That is why 
semantics is so damnably hard to do, and why people misunderstand 
each other and why even technical specs often fail to convey their 
intended meanings.

Let me ask you how you might feel if I were to dismiss the entire 
work of the TAG team by saying that the business of sending 
information over the WWW is all just a matter of sending bits and the 
rest can be "dealt with readily in human language", so what is all 
the fuss about. This is roughly equivalent to what you are saying to 
us here.  Speaking now as a member of other W3C WGs, this kind of 
response is so unhelpful as to be close to insulting.

>; for example, your admirably-clear assertion of what you'd like 
>mailto:jonathan@openhealth.org to mean. But today's Web doesn't care.

Maybe the machinery doesn't care - that is arguable.  But the success 
of the Web depends on more than just the machinery; and any kind of 
complete analysis of why the Web hums so well must need to take 
semantics and meaning into account.

>>>I don't understand the above.  URIs in the Web architecture are what
>>>they are, and what they are is effectively defined by a huge universe of
>>>deployed technology; they are character strings that can be used to look
>>>things up in databases and to retrieve representations.

Agreed.  Could you stop there? Why do you need to say that the 
retrieved representation is "of" a single "resource" which is 
"identified" by the URI and "part of" the WWW, and that this 
"resource" is constant and unchanging but also time-dependent and 
contextual? None of this seems like part of an *architectural* 
specification: it reads like an attempt to link the architecture to a 
semantic theory.  Unfortunately, considered as a nascent semantic 
theory, it is, as stated - and not to put too fine a point on it - 
complete nonsense.

>>>  This is the
>>>basis of the Web architecture.  If you want to use them as a basis for
>>>building the SW, that's fine, but please don't try to stop us from
>>>writing down an accurate description of reality as it empirically is.
>>Certainly reality includes RDF applications which date back to 1999? Are you
>>limiting your definition to HTTP servers? You ought not.
>As noted above, it is appropriate to attack the problem of writing 
>down the architectural bases of the SW world-view.  But today's Web 
>will continue to exist independent of SW superstructure, however 
>useful, and it's very important to capture the principles of its 


>Which I think the webarch doc is starting to do a good job of.

I wish it were true, but there is a gaping hole right at the center. 
The doc talks about resources and URIs identifying them, but nowhere 
says what this means. It also talks about representations, but 
apparently does not mean by that what that word usually means.

>>>We are *entirely* clear what URIs are in the context of the Web-that-is,
>>>and we have the software to prove it.  We are trying to write down that
>>>shared understanding.
>>He he he. If we were *entirely* clear what URIs where, why all the endless
>Because SW practitioners are trying to impose their world-view on a 
>reality that operates independently of it.

I am not trying to impose anything on anyone. I am only trying to 
understand what YOU are talking about.

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, 16 July 2003 15:46:24 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:55:59 UTC