RE: Uniform access to descriptions

Hello Pat,
________________________________
From: Pat Hayes [mailto:phayes@ihmc.us]
Sent: 08 April 2008 19:03
To: wangxiao@musc.edu
Cc: Williams, Stuart (HP Labs, Bristol); Jonathan Rees; www-tag@w3.org WG; Phil Archer
Subject: Re: Uniform access to descriptions

At 5:54 PM +0100 4/8/08, Xiaoshu Wang wrote:
Stuart,
Wrt to that resolution... a 303 response means *nothing*... if you happen to follow the redirection and find something useful about the thing you originally inquired of, that you trust and are prepared to stick in your reasoning engine, then you win - if not, of itself, the redirection has told you nothing/means nothing.

200 tells you that the response convey as representation of the (state of?) referenced thing.

If this is what TAG accepts, i..e, 200=*representation of* as oppose to "resource of".  I have no problem and would be happy with it.  My perception is that TAG is recommending either explicitly or implicitly the latter viewpoint.

Gentlemen, please both of you speak very slowly and carefully at this point, as a precise understanding here is critical.
Ok... treading very carefully and hoping not to trigger the land mines :-)

Stuart, did you mean that the response conveys a representation in the webarch sense of the referenced thing?
Yes I did.... in the sense that the response is a message and that the content of the message awww:represents the referenced thing.
 It would be helpful if every time the word 'represent' and its cognates are used in this very special sense, such usage were explicitly flagged, as it can very quickly lead to incomprehension when understood more broadly (as it is almost everywhere else in the English-speaking world.)
Ok, apologies. I havered over doing that and the form or the wording... I was try to avoid speaking of a class of things that might be spoken of as awww:Representation given discussions elsewhere about the utility of 'coining' a class of such things. I didn't quite manage it.
(Xiaoshu: from which it follows that in this case, the referenced thing in question must be something that has a webarch-representation; so, in this case, it cannot be some other kind of thing that cannot, by virtue of its very nature, have such a (webarch-)representation; so, to refer to such things - such, as we now might say, non-information resource things - requires something other than a 200 response. Thus goes the http-range-14 logic, as I understand it.
If it helps, I concur with you explaination of the logic.
 Note that in order to follow this, all we need to know is that there are things which (a) cannot have a representation in the webarch sense but (b) that we might wish to refer to with a URI. Their exact nature need not be specified, but I believe that the language of 'information resource' boils down to  an attempt to characterize this category of [things that cannot be webarch-represented by a byte stream].

Well yes... in the sense that the AWWW give definition of "information resource" is an attempt to characterize a category of [things that *can* be webarch-represented by a byte stream] which then gives rise to another category of things which cannot, could never be, so webarch-represented.
 And, centrally important, not having a representation in the webarch sense does not mean not having any kind of representation, being unrepresentable, or not being describable. The webarch sense of 'representation' is very specialized and narrow.)

Xiaoshu, what do you mean by "resource of" a referenced thing? In my understanding of the terminology, resources (and things) may have representations of them, but are not themselves "of" anything.

Pat

--


---------------------------------------------------------------------
IHMC               (850)434 8903 or (650)494 3973   home
40 South Alcaniz St.       (850)202 4416   office
Pensacola                 (850)202 4440   fax
FL 32502                     (850)291 0667    cell
http://www.ihmc.us/users/phayes      phayesAT-SIGNihmc.us
http://www.flickr.com/pathayes/collections
Thanks,

Stuart
--
Hewlett-Packard Limited registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England