Re: Resources and URIs

>>>It really would be most constructive if you'd suggest
>>>alternative text for the spec.
>>
>>Look, this is like God asking me to tell him what 'thou shalt not' 
>>means. How the hell do I know what text to suggest? I didnt invent 
>>the Web.  Here's the group, tasked with the job of writing the 
>>document which is supposed to define the basic terminology, and 
>>they want ME to tell them what they are talking about? That is 
>>crazy. [later: but see PS.]
>
>No, we want you to put up or shut up.  We want you to take the time to
>describe what it is you want the definition to say, because it is quite
>possible that you might come up with better wording than what we have now.
>That is standard operating procedure for editing specifications.

I have done, and you have rejected it.

I want the definition, if there must be one, to say that a resource 
can be anything.
If this is insufficient, I want it to say that it is anything that 
can be referred to.
However, my preference would be to say as little as possible, and the 
best way to do that is to say explicitly that no definition is being 
given, and to say why.

I have offered wording for all three cases.

I do not want the definition to use any words which can be 
interpreted as saying that a resource must be somehow identifiable or 
accessible.  I do not want the definition or the surrounding text to 
say anything which can be interpreted as meaning that URIs have 
unique associated resources, or that there is a fixed or unambiguous 
mapping from URIs to resources. If examples must be given of how URIs 
can identify resources, then they should be accompanied by a clear 
statement that they are not considered to be an exhaustive list of 
the possibilities, but are provided only as illustrative.

You know, this concept you are trying to define is awfully close to 
meaning 'anything'; and to ask for a precise and non-circular 
DEFINITION of a concept at that level of generality is almost certain 
to land you in conceptual confusion. Words like that don't have exact 
definitions.

>>I have never known what the word 'resource' is supposed to mean. I 
>>asked you once, and you told me to just forget about it and pretend 
>>it meant 'anything'; but I don't think that is what Roy and Tim 
>>Bray (or Tim B-L) have in mind. My problem is, I really cannot 
>>discover what it is that they do have in mind, and the more I read 
>>and listen, the worse it gets. There seem to be fundamental 
>>confusions built into the very suppositions of the discussion. The 
>>fact that the discussion has been going on for a decade doesn't 
>>make them any less confused.
>
>Oh, so you just want me to tell you what it means, and that will be the
>end of it?  Every time I do that, other people claim I am just arguing
>from authority.  Ask them if you want that kind of answer.

Well, you are the editor of the normative document which sets out to 
say what it means, right? So, yes, I do want you to say what it 
means, if you think there is a meaning. Of course, it can reflect the 
views of a larger bunch of people, but your name is on the cover. My 
name is on the cover of the RDF Semantics document, so I went to a 
lot of trouble  to try to make the intended meaning of all the 
technical jargon it in it as clear as I possibly could.  If people 
came back (as they did) and said that something was unclear or 
apparently didn't make sense, I tried to find ways of making it 
clearer, or put in prose to warn readers against misunderstandings. 
(I was careful to say that "resource" meant anything, for example, 
and that no boundaries were being placed on its meaning by RDF.)  If 
something seemed to be potentially misleading and wasn't really 
needed, I took it out. Its been through around fifty versions by now, 
I would estimate, and some parts have been re-written maybe a dozen 
times.

On the other hand, the parts of your document about which I 
originally asked for clarification, pointing out detailed apparent 
internal contradictions, are almost unchanged, reproduced verbatim 
from the original document with which you began; and your only 
response to request for clarification has been a series of brusque 
refusals to even consider that there could possibly be anything 
unclear. Frankly, I do not consider this to be professional behavior 
from a WG editor, regardless of any differences of opinion we may 
have.

BTW, I don't understand your attitude to this word "resource". You 
are insisting on DEFINING it, but now you refuse to SAY WHAT IT 
MEANS. Honestly , now, doesn't this seem to you to be rather a 
difficult position to put yourself in?

>>>If you can see
>>>more than one interpretation, suggest text for each
>>>and see which one gets more support. Or if you
>>>prefer one interpretation, suggest text to support
>>>that.
>>
>>I don't have any preferences, other than for someone to say what 
>>the hell they are talking about. We can adapt the model theories to 
>>just about anything, I reckon, as long as its defined well enough 
>>so that we can model it mathematically.  But I need something to 
>>work with (or a licence to define things myself.) I can't manage 
>>with a blurry notion of 'refers' which when examined a little 
>>closely turns out to be hiding arbitrary use/mention confusions, 
>>'social conventions', temporal indexicality, ambiguity of 
>>reference, confusions between retrieval and naming, frames of 
>>reference, individuation conditions which are not specified, and 
>>who knows what else - all implicit, but nowhere spelled out, in the 
>>apparent meaning of terminology described in a normative document - 
>>but which is also arbitrarily restricted in irrational ways, eg so 
>>as to make quantifiers meaningless.
>
>You obviously have something you want it to say, otherwise you would not
>be concerned about any of those things.

Well, I have my druthers, but I havn't been pressing them here. 
Really what I want most is for the current apparent contradictions to 
be resolved (or removed), and the meaning of the terminology being 
used made clear. That is what I asked for in the first message and 
have been asking for ever since, and you have consistently refused to 
even discuss the specific issues I have raised about the actual text. 
If 2396 finishes up defining "resource" too narrowly then RDF can 
just use a different word; that would be a pity, but at least we then 
all know where we stand.

>  In any case, the definitions
>aren't in the normative part of the document anyway.  They are attempts
>to answer the most commonly asked questions.
>
>>>Neither Roy nor anybody else is in a good position
>>>to answer the sort of clarification questions you're
>>>asking.
>>
>>Well, then, if that is really true, then they are not up to the 
>>job, and should resign or find someone who is in a position to do 
>>it. I don't mean to be rude, but this is like telling me that 
>>there's nobody on the XML Schema group who knows anything about XML.
>
>I didn't know I had the job of telling everyone the definitions.
>Heck, if that's the case, I don't need a working group.

I was referring to the working group, hence the use of the plural.

>The definitions I like are already in the document.  I already gave
>you a reference that defines them as completely as I wish them to be
>defined.

I must have missed that somewhere. What reference?

>  If that isn't good enough for you, how am I supposed to
>know what will be good enough?

Well, you could read my explanations of what it is about them that 
seems unclear or self-contradictory, and try to respond to them more 
helpfully than (1) you are wrong (2) you are ignorant (3) Ive already 
said it and its perfectly clear to me, go away (4) you are confused 
(I knew that already)  (5) you are confused because you are too much 
of a mathematician  (6) you are confused because you are too much of 
a philosopher or (7) you are from the Semantic Web and we don't give 
a f**k about the Semantic Web here.

Apparently you have already, for example, diagnosed that part of my 
problem is that I am using the wrong notion of "identity".  But you 
haven't told me what the RIGHT notion of 'identity' is.

>As near as I can tell, the entire source of your confusion is due to
>assuming that the word "identity" refers to the identity function of
>mathematics, which as I said at the beginning of this thread is not
>the case.

Not just mathematics, but also all of philosophical logic, formal 
semantics, a large part of semiotics and linguistics, all of computer 
science and most of science more generally, virtually all software 
reasoning engines and also in quite a lot of normal English usage, in 
fact.  But OK, fair enough, its your document: but then you need to 
SAY what sense it IS being used in.

I know I am a comparative newbie in Web circles, but I honestly think 
I am a reasonably careful reader and not a complete idiot, and I have 
relevant background competence in a fair number of technical areas as 
well as a smattering of computer science.  In particular, Ive worked 
in many interdisciplinary settings where one has to explain things to 
people who use a different jargon. And I have been trying as hard as 
I know how to make sense of this text now for about two years, have 
had long detailed discussions with many people in the W3C from Tim 
Berners-Lee on down about what it means, and have been deeply 
involved in many debates arising in trying to get formal semantics 
done in a Web context while trying to conform to what it says. I'm 
trying hard here, and I have to tell that what you say in the 
relevant passages simply does not make sense to me.  They seem to 
have internal contradictions, to be based on unstated assumptions 
with which I am unfamiliar, to use what is apparently technical 
terminology in ways that are unfamiliar, and they don't explain any 
of this.  And the more we have this discussion, the less sense it 
makes.  Doesn't the fact that I am having such trouble might at least 
hint at the possibility, if I can make the suggestion without hubris, 
that there might be some real explanation work still to be done here?

Now, as I said, if there is some background area of expertise here 
about which I am ignorant, please let me know; though it might also 
be helpful to also refer to it in the document, by the way, if only 
in a glossary.

I am honestly puzzled by your reaction to all this, by the way. If 
someone says to you, what did you mean? Surely your proper response 
is not, you tell me what I ought to have said. If they knew that then 
they wouldn't be asking.  And it doesn't seem very helpful to just 
say the same thing back again, either, or to observe that it seemed 
clear enough to you; in fact, that last response is downright 
insulting.  I think that your (and Dan Connolly's ) reaction might be 
predicated on the assumption that *everyone* knows what 'resource' 
means, really, so of course my questioning the definition is just 
rhetorical, and I ought to cut the crap and help with the 
wordsmithing. But really, I don't know.  What is being said about 
resources doesn't make any overall coherent sense, unless 'resource' 
just means 'anything'; but it is clear that isn't what you mean.  And 
I can report that in my experience, I don't think that many other 
people know, either.  Or at any rate, the ones who think they know 
don't all agree with one another. So, at the risk of repetition, I 
can't 'put up' because I don't know what the hell y'all are talking 
about.

I'm not here saying that the Web isn't real or that URIs don't work 
properly, or whatever. I know that URIs work very well indeed. What 
doesn't make sense is what people SAY about them.  And one of the 
reasons it doesn't make sense  is because it doesn't seem to fit with 
all the ways I see URIs actually being used on the Web, in fact; and 
it certainly doesn't fit with all the ways that URIs are being used, 
and are planned to be used, on the semantic web.

>I can clarify that in the text, assuming I ever get a chance
>to work on the document.
>
>>"This document specifies the syntax of URIs, which are a form of 
>>global identifier used in Web protocols and languages.  Particular 
>>uses of URIs, and their intended meanings in various contexts, are 
>>described in other specifications. In general, the entities 
>>referred to or identified by URIs when used in Web contexts are 
>>called "resources"., but this document does not specify the nature 
>>of resources or to restrict resources to any particular category of 
>>entities."
>>
>>and leave it at that.  Nothing else at all about resources, no 
>>examples, no discussion.
>
>No.  Look, you guys aren't the ones who have to answer questions in the
>absence of definitions.  I do.

Well, just say that its up to other specs to determine what they 
consider to be resources.  Not your job, you just do the URI syntax. 
Seems fair to me.  We've taken very much that attitude with RDF 
semantics. There are several aspects which the semantics carefully 
avoids specifying a meaning for, and even boasts about it. I've had 
protests, and the reply is: sorry, whatever we said here would cause 
someone to complain that it would make their deployed code 
non-compliant. We can't keep everyone happy, so we aren't saying 
anything, and applications can make their own calls about it. (There 
is a 'preferred' meaning, and we say what it is and recommend it; but 
we don't formalize it.) This could lead to some interoperability 
problems; but they already exist anyway, that's why we have the 
issue.  If they matter, people will work out ways around them and the 
best man will win, and maybe then in the next version of RDF another 
WG will set that in stone.  But we don't try to put the world to 
rights before its ready to cure itself.

(You wouldn't want to be telling XHTML what it ought to do with URIs, 
or to tell HTTP what it ought to do, etc., right? So why do you want 
to tell RDF what kind of thing its URIs have to refer to? )

>I refuse to leave what has been deployed
>in an unspecified state, regardless of how many arguments that causes
>in the Semantic Web.

So 'anything that has an identity' is an adequate specification? Come 
on, be serious.

It seems to me that these arguments have been going on long before 
the semantic web came along. The difference is that the Sweb needs to 
be precise about these things in a new way, so what you say about 
identification and meaning now has potential technological 
consequences, which is why I'm getting so alarmed. Its not enough to 
just specify it in an armchair-lexicographer kind of way: if you 
specify it, you need to get it right, and there are technical 
criteria for what counts as right; getting it wrong will actually 
break things.

>If SemWeb needs a better definition, then its
>proponents can reach consensus

'proponents'? You make it sound like some political slogan. You are 
talking about many hundreds of people spread all around the planet, 
at least three W3C working groups, large US/European consortia and 
government and industrial standards groups concerned with a 
technology which has been in mature industrial use for about 15 years 
now.  And anyway, your attitude here seems strange, almost hostile. 
Don't you think that the semantic web is part of the web, and that 
you therefore have some responsibility towards it, if only as one 
user among many? Or do you only represent the 'old' web?

>on what it should be and provide me
>with an appropriate text that has no adverse impact on deployed
>implementations of URI.

In what way does my proposed wording adversely impact any 
implementations, deployed or otherwise? It was carefully phrased so 
as not to impact *anything*.

>I am not even remotely confused about what resource means.

I'm glad to hear it; but as the editor of the document, don't you 
think that it might be worth trying to enlighten the rest of us? 
Surely, the task of the document is to tell everyone else what is 
meant. I have given you several lists of examples which you could use 
to help clarify your meaning

>  That does
>not mean the definition can't be improved.

About the only changes that I can think of are the ones I have 
already suggested.

I would urge you to reconsider the 'minimalist' option in the PS of 
my last message - the exact words don't matter, but the idea -  as it 
isn't going to cause any actual harm or effect anyone's definition of 
"resource",  and it can be used to deflect future criticism.  I could 
rewrite the entire introduction in this style, with some of Tim 
Bray's material worked in, if you think it might be helpful.

Pat

-- 
---------------------------------------------------------------------
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@ai.uwf.edu	          http://www.coginst.uwf.edu/~phayes
s.pam@ai.uwf.edu   for spam

Received on Thursday, 24 April 2003 02:15:50 UTC