Re: Another whenToUseGet-7 counter-proposal

On Tue, Apr 23, 2002 at 11:04:38PM -0700, David Orchard wrote:
> So when the fateful question is asked, here's some suggested wordings for
> how to deal with SOAP 1.2 and the wsarch group.
> "The TAG finds that there SHOULD be a usage scenario in the web services
> architecture, specifically there SHOULD be a standardized mechanism for
> using HTTP GET to retrieve some SOAP messages, particularly for
> browser/hyper-link traversal applications.  The TAG suggests to the Web
> Services Architecture Working Group that it SHOULD add such a usage scenario
> in it's efforts.  Further, the TAG believes there SHOULD a simple solution
> that hits an 80/20 mark.  The TAG believes a simple solution SHOULD be
> treated as a high priority item for the wsawg in it's efforts."

I have a better proposal for you.  We have already addressed this issue
in the XMLP WG;

The resolution to that should suffice for the TAG, IMO.

As I said to Anne on www-talk;

  "Also, as you know, I've worked since day one in the XML Protocol WG to
   ensure that SOAP *does* have a use that is compatible with Web
   architecture.  So I sort of disagree with Roy when he says that SOAP
   itself is bad.  But on the other hand, if nobody uses it in this way,
   then at what point do you say that the technology itself is flawed?
   I don't know, but I have no intention of voting against SOAP 1.2 because
   of the considerable resources that my company expended in making sure
   that it could be used in this way."

You don't need a GET binding of SOAP in order to use SOAP in a Web
friendly manner, you just need to *also* use GET (and PUT and DELETE,
where appropriate) by itself.  For example, I could send a SOAP message
with HTTP POST to a resource, and then invoke GET on that same resource
to observe the state change induced by the POST.

Mark Baker, Chief Science Officer, Planetfred, Inc.
Ottawa, Ontario, CANADA.

Received on Wednesday, 24 April 2002 08:32:59 UTC