Re: Second Draft of summary of TAG issue abstractComponentRefs-37

Jonathan Borden writes:

> URIs should be considered opaque by processes that
> assign URIs to resources. URIs should only be parsed by
> processes that retrieve representations of resources
> identified by a particular URI a.k.a. the process of
> "URI resolution"

This seems to me to be an over simplification.  How, for example, do we 
explain the query component ("?") of a URI and its incremental 
construction by the GET mode of HTML forms?  [1].  RFC 2396 is very clear 
that the query string is part of the URI, and surely that string is not 
best thought of as opaque at either the client or the server according to 
this common usage.   To just say "URIs should be opaque to processes 
assigning URIs to resources" seems to ignore not only query strings, but 
also the provision of a hierarchical idiom in 2396 and its common (and 
presumably intended) application to hierarchical resources such as 
filesystems.

I think the opacity principle is very important, but I think the TAG has a 
real opportunity to set it forth clearly and in a form that will answer 
practical questions about its application to real-world scenarios.  I've 
requested this of the TAG before [2], and at the risk wearing out my 
welcome, Jonathan's note prompts me to repeat that suggestion.  Any chance 
the TAG would want to open an issue on clarifying the opacity principle? 
Thank you.

Noah

[1] http://www.w3.org/MarkUp/html-spec/html-spec_8.html#SEC8.2.2
[2] http://lists.w3.org/Archives/Public/www-tag/2003Jan/0158.html

------------------------------------------------------------------
Noah Mendelsohn                              Voice: 1-617-693-4036
IBM Corporation                                Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------

Received on Monday, 28 April 2003 15:59:04 UTC