W3C home > Mailing lists > Public > www-ws-desc@w3.org > July 2004

RE: Issue 189 proposals:

From: David Orchard <dorchard@bea.com>
Date: Tue, 27 Jul 2004 11:48:11 -0700
Message-ID: <32D5845A745BFB429CBDBADA57CD41AF093330CA@ussjex01.amer.bea.com>
To: "Hugo Haas" <hugo@w3.org>
Cc: "Sanjiva Weerawarana" <sanjiva@watson.ibm.com>, <www-ws-desc@w3.org>

We made a decision, which I disagreed with but am still trying to live
with in spirit and in wording, that the URI could contain information
that was in the body, but the URI could not contain information that was
not in the body.  

The 2 cases where part of the data that is in the body may also need to
go into the URI, but the entire element should be serialized, are PUT
and POST, where there is a body and serialization into the URI.  The
point is that there are 2 aspects of serialization for these: the
serialization into the body and into the URI.

GET and DELETE have only a serialization into the URI, but they can also
do with truncating the instance data.  In this case, then the same type
could be used for GET/PUT/POST/DELETE operations.

Imagine where there is an Artist with an ID element, and the ID element
is the unique identifier (and damn it, we should allow attributes for
this case!!).
<Artist><ArtistID>5</ArtistID><stuff/></Artist>

I may want to PUT a new artist using ArtistID 5, ie PUT /Artist/5.  This
is something like location="/Artist/{ArtistID/}.  Same holds true for
POST, where the client could be suggesting that /Artist/5 is the URI for
the new artist.  The body of these would be the entire Artist element.

I may want to GET the artist or delete the artist, and the same location
value would hold true.

I hope this clears up any confusion.

Cheers,
Dave




-----Original Message-----
From: Hugo Haas [mailto:hugo@w3.org] 
Sent: Tuesday, July 27, 2004 9:44 AM
To: David Orchard
Cc: Sanjiva Weerawarana; www-ws-desc@w3.org
Subject: Re: Issue 189 proposals:

* David Orchard <dorchard@bea.com> [2004-07-26 14:44-0700]
> Hugo has already found his answer, which hopefully is that the rest of
> the xml is not serialized at all.

Oh, I got it wrong then. I wrote into the spec:

  All elements of the instance data from the input message NOT cited
  by the location attribute information item are serialized in the
  message body following the rules of the "application/xml" (see 3.8.2
  Serialization as application/xml ).

at:

 
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-bindings.
html?rev=1.57&content-type=text/html;%20charset=utf-8#_http_operation_lo
cation_notcited_body

with an example I added at:

 
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl20/wsdl20-bindings.
html?rev=1.57&content-type=text/html;%20charset=utf-8#urlencoded_example
_body

Hmmm... I had understood that it was an interesting proposal for PUT
or POST, and therefore that something needed to go into the message
body.

But your proposal doesn't populate the body then, just limits the
serialization in the request URI. Hmmm... now I'm confused about its
use.

In any case, I'll let you a chance to see what I had written to
confirm that I need to change it before I change the draft.

Regards,

Hugo

-- 
Hugo Haas - W3C
mailto:hugo@w3.org - http://www.w3.org/People/Hugo/
Received on Tuesday, 27 July 2004 14:48:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:32 GMT