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

Re: Issue 169: Propose http method in the operation interface to simplify http binding.

From: Sanjiva Weerawarana <sanjiva@watson.ibm.com>
Date: Tue, 13 Jul 2004 09:37:56 +0600
Message-ID: <044c01c4688a$cab5bca0$f24e4109@LANKABOOK>
To: <www-ws-desc@w3.org>

IIRC we added @safe as a way to satisfy the TAG. I for one did not
(and do not) believe it belongs in the interface level (because
of HTTP specificness) and don't accept that as a trojan horse to
add more HTTPisms to the interface. I let it go because I'm not
convinced many people will use it and its use is optional in any
case.

How about the following: IMO there's nothing wrong with a binding
in WSDL choosing to add properties to abstract components. That is,
I don't believe we say anywhere that a binding can only add stuff
inside the <binding> element. So, if you really want to add the
"web method" concept to the interface, then add it as:

    <operation name=".." safe="yes|no" whttp:webMethod="whatever">
        ...
    </operation>

If you want to add it as a feature that's fine too; both are forms
of extensibility. I still am not at all convinced that the concept
is abstract and belongs in the interface for all bindings, but if
the HTTP binding wishes to define something that can be asserted
at the interface level that's ok with me.

We would still need to write into the SOAP binding something like
Hugo's rules for the permissible values for this attribute when
the binding is SOAP. I assume someone else will worry about the
HTTP binding's implications (maybe remove @whttp:method?).

Sanjiva.

----- Original Message ----- 
From: "David Orchard" <dorchard@bea.com>
To: "Jeffrey Schlimmer" <jeffsch@windows.microsoft.com>; "Amelia A Lewis"
<alewis@tibco.com>
Cc: <www-ws-desc@w3.org>
Sent: Tuesday, July 13, 2004 2:53 AM
Subject: RE: Issue 169: Propose http method in the operation interface to
simplify http binding.


>
> Does the latter principle apply outside of WSDL, that is to suggest that
third parties should not extend wsdl:interface/* unless it maps to (nearly)
all bindings?
>
> Doesn't that latter principle also say that "safe" should be moved into
the binding?  For example, SMTP doesn't have a "safe" operation as defined
by RFC 2616.  It's got "MAIL", "RCPT", "DATA", etc.
>
> Also, the motivation for "safe" is primarily HTTP specific but potentially
useful for other protocols.  "safe" was added explicitly to give greater
support to HTTP by allowing applications to indicate when bound to HTTP, a
GET operation may be used.  Which is the same motivation for webMethod at
the interface operation level, allowing applications to indicate when bound
to HTTP, a particular opertion may be used.  Not much difference in my mind.
>
> Cheers,
> dave
>
> > -----Original Message-----
> > From: Jeffrey Schlimmer [mailto:jeffsch@windows.microsoft.com]
> > Sent: Thursday, July 08, 2004 11:29 PM
> > To: Amelia A Lewis; David Orchard
> > Cc: www-ws-desc@w3.org
> > Subject: RE: Issue 169: Propose http method in the operation interface
> > to simplify http binding.
> >
> >
> > If something is specified in more than one binding, will the
> > WG move it
> > up into wsdl:interface?
> >
> > Or should the scope of wsdl:interface/* cover only that which is
> > expected to be utilized by (nearly) all bindings?
> >
> > Applying the latter principle would suggest that webMethod does not
> > belong on wsdl:interface/* because it would not map to all bindings.
> >
> > > -----Original Message-----
> > > From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org]
> > On
> > > Behalf Of Amelia A Lewis
> > > Sent: Tuesday, June 29, 2004 12:51 PM
> > > To: David Orchard
> > > Cc: www-ws-desc@w3.org
> > > Subject: Re: Issue 169: Propose http method in the
> > operation interface
> > to
> > > simplify http binding.
> > >
> > >
> > > On Tue, 29 Jun 2004 12:40:48 -0700
> > > David Orchard <dorchard@bea.com> wrote:
> > > > The changes to WSDL are:
> > > > 1. WSDL interface operations contain optional webMethod attribute.
> > > > This is an HTTP operation name.
> > >
> > > Strongly -1.  We have worked hard to separate keep the abstract
> > interface
> > > abstract.  HTTP methods are binding-specific, not interface-level
> > > abstractions.
> >
Received on Tuesday, 13 July 2004 00:34:07 GMT

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