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

Interacting with a URI, when is the result not a representation and what is it called.

From: David Orchard <dorchard@bea.com>
Date: Wed, 5 Feb 2003 12:27:21 -0800
To: <www-tag@w3.org>
Message-ID: <000801c2cd54$fd79f390$940ba8c0@beasys.com>

We've been having some discussions on the xml-dist-app mailing list on the
relationship between REST, the Web, and SOAP.  I'm not trying to get into
any kind of criticism or debate about whether things are on the web, not on
the web, etc.  I simply want to come up with some terms and relationships
that we all can agree to.  I've been trying to come up with a definitive
relationship between SOAP messages, infosets, representations, and what the
media-type actually identifies in the context of SOAP.

As interacting with URIs in a RESTful manner - that is dereferencing URIs -
can result in a representation, what is the verb/result of non-restful
interactions with URIs?  For example, say an RPC style invocation that
tunnels over HTTP POST, or a non-RESTfully designed HTML form that embeds a
method in it.  What are the interactions/results called? [1].

If dereferencing URIs always results in representations [2], and yet some
SOAP interactions don't produce representations [3], then clearly the SOAP
interactions that don't produce representations aren't "dereferencing".
There's a verb missing from our lexicon.

Elaborating from SOAP to a non-Restful HTML Page [4], what is the result of
interacting in a form that is non-restfully designed.

My Question #1 is thus: It seems like there are two choices wrt
interactions:
a) Interacting with a URI always produces a representation, whether or not
it's a badly designed interaction or not.  That is, representations are the
complete set of results from the interaction.
b) Interacting with a URI can result in something other than a
representation.  This means we probably need another noun for what is the
result of the interaction. And we also need another verb for the name of the
type of interaction.

Which leads to question #2, if the answer to #1 is b, then are those URIs -
such as a bad HTML form - considered "on the web" or not?

I can only conclude the answer to #2 is yes, that is even badly designed
non-RESTFUL forms are "on the web".  In which case, we need to talk about
representations AND something else in describing the web architecture.
Which has distinct implications for our arch document and the editing that
we are going to do over the next few days.

Cheers,
Dave

[1] http://lists.w3.org/Archives/Public/xml-dist-app/2003Feb/0040.html
[2] http://lists.w3.org/Archives/Public/xml-dist-app/2003Feb/0035.html
[3] http://lists.w3.org/Archives/Public/xml-dist-app/2003Feb/0034.html
[4] http://lists.w3.org/Archives/Public/xml-dist-app/2003Feb/0032.html
Received on Wednesday, 5 February 2003 15:29:29 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:16 GMT