W3C home > Mailing lists > Public > public-awwsw@w3.org > May 2008

RE: Use cases from Jonathan

From: <noah_mendelsohn@us.ibm.com>
Date: Tue, 6 May 2008 21:24:35 -0400
To: "Booth, David (HP Software - Boston)" <dbooth@hp.com>
Cc: Jonathan Rees <jar@creativecommons.org>, "public-awwsw@w3.org" <public-awwsw@w3.org>
Message-ID: <OF022B3255.4023769E-ON85257442.0006F303-85257442.0007A5BA@lotus.com>

David Booth writes:

> We know that conneg is for conveying subsets of the same information

I think I'm right that one common use case is for different natural 
language renderings of the "same" abstract document, e.g. Greek, Chinese, 
French and English versions of the same press release.  I'm not sure that 
calling those "subsets of the same information" quite captures this 
particular use case.  My view is that, in this case, there is an abstract 
press release in the "mind" of whoever owns the URI that's supporting the 
conneg.  Perhaps that is viewed as being fundamentally in, say, English, 
with the other versions being viewed as approximations in the other 
languages.  Perhaps no one natural language version is considered by the 
owner to be definitive, in which case presumably the measure of fidelity 
is whether each of the translations successfully conveys whatever abstract 
points the owner of the resource considered essential to the press 

I can imagine at least two ways of modeling this use case:  1) the 
resource is the union of all the translations -- I don't like this because 
I posited that this was not how the owner of the resource preferred to 
think about it or 2) there is one abstract resource, and the various 
representations capture it with greater or lesser fidelity, much as an 
image/jpeg might be a lower fidelity representation of an image that is 
captured with higher fidelity in an image/gif.  Anyway, if you buy version 
(1), then I think you get to say that we are conveying subsets of the 
state of the resource, but I don't think you get to call it "subsets of 
the >same information<".  The various translations are ultimately 
different translations.  Or, if you go with option #2 (which I prefer), 
you get to say that what is being conveyed in each case is a 
representation of the same information,  but with varying degrees of 
fidelity.  I don't think the word "subset" applies to model #2.


Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
Received on Wednesday, 7 May 2008 01:24:18 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:21:06 UTC