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

Re: WhenToUseGet-7 comments

From: Dan Connolly <connolly@w3.org>
Date: Wed, 17 Sep 2003 08:43:41 -0500
To: David Orchard <dorchard@bea.com>
Cc: www-tag@w3.org
Message-Id: <1063806221.5534.297.camel@dirk.dm93.org>

On Mon, 2003-09-15 at 13:45, David Orchard wrote:
> At the end of the draft finding, it says "Section 3 WSDL 1.2 Bindings [WSDL]
> provides a binding to HTTP GET, which makes it possible to respect the
> principle of using GET for safe operations. However, to represent safety in
> a more straightforward manner, it should be a property of operations
> themselves, not just a feature of bindings.".
> 
> I'm not quite sure what this means.

As I mentioned in the telcon, we hear a certain amount about folks
wanting to use POST all the time because it's supported better
by point-and-click development tools.

My goal is that these point-and-click development tools include
an "[ ] safe operation" checkbox next to each operation. The
results of that checkbox go into the WSDL file as a property
of the operation, and the software that binds it to HTTP would
use GET, provided there are, say, less than 8 parameters
that are all scalars.



>   Does this suggest that there should be
> a general property of operations that specifies safetey, which could then be
> bound to POST?

Marking HTTP POST operations safe is different from what I had in mind.
I don't see anything wrong with it, but I don't see very much benefit
from it either, compared to the benefit of making the results
of safe operations addressable.


>   That is a operation declarated safe could be bound to POST?
> I think that's the right interpretation, in which case I'm comfortable with
> the wording.
> 
> Interestingly, I just heard the need from 2 customers last week who want to
> be able to have a "ping" function to test the availability of a service.
> 
> At this point, I don't think the WSDL WG has accepted this requirement.  The
> editors copy of the requirements document (last updated feb 2003), has
> requirement R125 as the only TAG sourced requirement.
> "R125
> The normative description of the InterfaceBinding for SOAP 1.2 MUST support
> the SOAP 1.2 MEP for HTTP GET in and HTTP SOAP out. (From TAG. Last
> discussed 26 Sep 2002.)
> "
> 
> I propose to ask them for some clarification and call the finding to their
> attention.

Yes, as we discussed, that seems like a good idea.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Wednesday, 17 September 2003 09:43:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:20 GMT