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

RE: Proposed resolutions for issues 146 and 150

From: Arthur Ryman <ryman@ca.ibm.com>
Date: Wed, 24 Mar 2004 17:03:08 -0500
To: "Martin Gudgin" <mgudgin@microsoft.com>
Cc: "Jacek Kopecky" <jacek.kopecky@systinet.com>, "Tom Jordahl" <tomj@macromedia.com>, "WS Description List" <www-ws-desc@w3.org>, www-ws-desc-request@w3.org
Message-ID: <OF4319AB56.B5F6087D-ON85256E61.00780EC4-85256E61.00792147@ca.ibm.com>
Gudge,

I think that's the wrong conclusion. The message is abstract at the 
interface definition level. So if someone wanted to describe SOAP messages 
that had multiple elements then they could define a suitable binding. 
Isn't this analogous to the HTTP binding and URL encoding of parameters? 
Clearly there is no element there. The HTTP binding pulls apart the 
content of the element and stuffs it into a URL.

So the real question is: Do we want WSDL to describe messages so that SOAP 
bindings are simple to describe, or do we want to preserve the concept of 
abstract messages (based on a single element) and force the heavy lifting 
to be done in the bindings?

Arthur Ryman,
Rational Desktop Tools Development

phone: +1-905-413-3077, TL 969-3077
assistant: +1-905-413-2411, TL 969-2411
fax: +1-905-413-4920, TL 969-4920
mobile: +1-416-939-5063
intranet: http://w3.torolab.ibm.com/DEAB/



"Martin Gudgin" <mgudgin@microsoft.com> 
Sent by: www-ws-desc-request@w3.org
03/23/2004 06:53 PM

To
"Jacek Kopecky" <jacek.kopecky@systinet.com>, "Tom Jordahl" 
<tomj@macromedia.com>
cc
"WS Description List" <www-ws-desc@w3.org>
Subject
RE: Proposed resolutions for issues 146 and 150







Jacek,

It seems odd ( to me at least ) that WSDL not allow me to describe
messages that are clearly OK per the SOAP spec.

Gudge

-----Original Message-----
From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org] On
Behalf Of Jacek Kopecky
Sent: 23 March 2004 06:20
To: Tom Jordahl
Cc: 'WS Description List'
Subject: RE: Proposed resolutions for issues 146 and 150


Tom, I originally meant the issue 146 as really allowing anything in the
message, but I will have no problem with constraining that to "any
single element", it suits the usecase I have in mind here - a
content-based router endpoint that receives any message.

This way the spec will be clearer and more consistent and the
restriction to a single element in SOAP Body doesn't seem too bad; noone
knows how to handle multiple elements there anyway. 8-)

                   Jacek Kopecky

                   Systinet Corporation
                   http://www.systinet.com/




On Mon, 2004-03-22 at 23:07, Tom Jordahl wrote:
> I actually never believed we were discussing (4), I had always assumed
(3).
> I am also against the idea that you can get away with sticking
*anything* in
> to the message.  Now I understand why Umit is so worked up. :-)
> 
> I propose we clarify the meaning of "#any" to be explicit that we are
> specifying "any element", not "any stuff you want".
> 
> 
> --
> Tom Jordahl
> Macromedia Server Development
> 
> -----Original Message-----
> From: Roberto Chinnici [mailto:Roberto.Chinnici@Sun.COM] 
> Sent: Monday, March 22, 2004 4:36 PM
> To: Arthur Ryman
> Cc: Sanjiva Weerawarana; Martin Gudgin; Tom Jordahl; WS Description
List;
> www-ws-desc-request@w3.org
> Subject: Re: Proposed resolutions for issues 146 and 150
> 
> I find the current syntax nice and readble in three of the four cases:
> 
>    1)  element="myns:Foo"
>    2)  element="#none"
>    3)  element="#any" (where "#any" means "any element")
> 
> It's the fourth case, i.e.
>    4)  element="#any" (where "#any" means "anything, any kind of
content")
> that is problematic.
> 
> I'm actually having second thoughts on conflating (3) and (4).
> 
> I think that Umit has a point when she says that by adopting (4) we've
> moved away from an element-based content model representation.
> 
> Moreover, given that some bindings might have restrictions on the
> allowable payloads for a message, it seems important to distinguish
> between (3) and (4). Otherwise an application written to the abstract
> layer of WSDL will feel authorized, upon encountering a message
> definition which specified element="#any", to pass arbitrary content
> around, including content of a kind that will be systematically
rejected
> by the binding in use. Then we'd fall back again in the trap of
writing
> applications to a specific binding rather than to the abstract
interface.
> 
> Roberto
> 
> 
> Arthur Ryman wrote:
> > 
> > Sanjiva,
> > 
> > The attribute @element formerly refered to the QName of an element 
> > (GED). However, now it may not refer to an element. In fact, the
message 
> > content might be a simple type, or anything else, including nothing.
So 
> > it is a minor misnomer to call the attribute @element. However, most
of 
> > the time it will refer to an element. Logically, the attribute
describes 
> > the message content, which is often, but not always, an element.
> > 
> > Arthur Ryman,
> > Rational Desktop Tools Development
> > 
> > phone: +1-905-413-3077, TL 969-3077
> > assistant: +1-905-413-2411, TL 969-2411
> > fax: +1-905-413-4920, TL 969-4920
> > mobile: +1-416-939-5063
> > intranet: http://w3.torolab.ibm.com/DEAB/
> > 
> > 
> > *"Sanjiva Weerawarana" <sanjiva@watson.ibm.com>*
> > Sent by: www-ws-desc-request@w3.org
> > 
> > 03/16/2004 10:02 PM
> > 
> > 
> > To
> >              "Martin Gudgin" <mgudgin@microsoft.com>, "Tom Jordahl" 
> > <tomj@macromedia.com>, Arthur Ryman/Toronto/IBM@IBMCA
> > cc
> >              "WS Description List" <www-ws-desc@w3.org>
> > Subject
> >              Re: Proposed resolutions for issues 146 and 150
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > I'm confused .. I thought we're talking about special values to
> > assign to the operation/(input|output)/@element attribute to
> > indicate any content (#any) or no content (#empty).
> > 
> > What does this have to do with changing the name of the attribute?
> > 
> > Sanjiva.
> > 
> > ----- Original Message -----
> > From: "Martin Gudgin" <mgudgin@microsoft.com>
> > To: "Tom Jordahl" <tomj@macromedia.com>; "Arthur Ryman"
<ryman@ca.ibm.com>
> > Cc: "WS Description List" <www-ws-desc@w3.org>
> > Sent: Wednesday, March 17, 2004 1:43 AM
> > Subject: RE: Proposed resolutions for issues 146 and 150
> > 
> > 
> > Have you implemented it already? ;-)
> > 
> > Gudge
> > 
> > P.S. I've always thought it mildly amusing to have an attribute
whose
> > name is element ( or vice versa ) ;-)
> > 
> > 
> > ________________________________
> > 
> > From: www-ws-desc-request@w3.org
> > [mailto:www-ws-desc-request@w3.org] On Behalf Of Tom Jordahl
> > Sent: 16 March 2004 11:01
> > To: 'Arthur Ryman'
> > Cc: 'WS Description List'
> > Subject: RE: Proposed resolutions for issues 146 and 150
> > 
> > 
> > 
> > 
> > 
> > We just changed the name of this attribute to "element".
> > 
> > -1 to changing it AGAIN.
> > 
> > 
> > 
> > --
> > Tom Jordahl
> > Macromedia Server Development
> > 
> > -----Original Message-----
> > From: Arthur Ryman [mailto:ryman@ca.ibm.com]
> > Sent: Tuesday, March 16, 2004 1:05 PM
> > To: Tom Jordahl
> > Cc: 'Jonathan Marsh'; 'WS Description List';
> > www-ws-desc-request@w3.org
> > Subject: RE: Proposed resolutions for issues 146 and 150
> > 
> > 
> > 
> > 
> > Correction to my note:
> > 
> > s/elementReference/element/
> > 
> > Same comment applies. It's not an element anymore.
> > 
> > Arthur Ryman,
> > Rational Desktop Tools Development
> > 
> > phone: +1-905-413-3077, TL 969-3077
> > assistant: +1-905-413-2411, TL 969-2411
> > fax: +1-905-413-4920, TL 969-4920
> > mobile: +1-416-939-5063
> > intranet: http://w3.torolab.ibm.com/DEAB/
> > 
> > 
> > 
> > Tom Jordahl <tomj@macromedia.com>
> > Sent by: www-ws-desc-request@w3.org
> > 
> > 03/16/2004 09:30 AM
> > 
> > To
> > 
> > "'Jonathan Marsh'" <jmarsh@microsoft.com>, "'WS Description List'"
> > <www-ws-desc@w3.org>
> > 
> > cc
> > 
> > 
> > 
> > Subject
> > 
> > RE: Proposed resolutions for issues 146 and 150
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > Jonathan,
> > 
> > You meant to say "elementReference is the name of a type so it
> > would NOT appear in the syntax"
> > 
> > Right?
> > 
> > 
> > --
> > Tom Jordahl
> > Macromedia Server Development
> > -----Original Message-----
> > From: www-ws-desc-request@w3.org
> > [mailto:www-ws-desc-request@w3.org] On Behalf Of Jonathan Marsh
> > Sent: Monday, March 15, 2004 4:48 PM
> > To: WS Description List
> > Subject: RE: Proposed resolutions for issues 146 and 150
> > 
> > elementReference is the name of a type so it would appear in the
> > syntax.  I like messageBody better too.  Or I suppose we could just
get
> > rid of the reference altogether, right?
> > 
> > <xs:attribute name="element" >
> >       <xs:simpleType>
> >               <xs:union memberTypes="xs:QName">
> >                       <xs:simpleType>
> >                               <xs:restriction base="xs:token">
> >                                       <xs:enumeration
> > value="#any" />
> >                                       <xs:enumeration
> > value="#empty" />
> >                               </xs:restriction>
> >                       </xs:simpleType>
> >               </xs:union>
> >       </xs:simpleType>
> > </xs:attribute>
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > ________________________________
> > 
> > 
> > 
> > From: Arthur Ryman [mailto:ryman@ca.ibm.com]
> > Sent: Monday, March 15, 2004 12:58 PM
> > To: Sanjiva Weerawarana
> > Cc: Jacek Kopecky; Jonathan Marsh; WS Description List;
> > www-ws-desc-request@w3.org
> > Subject: Re: Proposed resolutions for issues 146 and 150
> > 
> > 
> > Sanjiva,
> > 
> > The XML Schema looks fine except for a couple of minor errors.
> > Here's a corrected version:
> > 
> >       <xs:attribute name="element" type="elementReference" />
> > 
> >       <xs:simpleType name="elementReference">
> >               <xs:union memberTypes="xs:QName">
> >                       <xs:simpleType>
> >                               <xs:restriction base="xs:token">
> >                                       <xs:enumeration
> > value="#any" />
> >                                       <xs:enumeration
> > value="#empty" />
> >                               </xs:restriction>
> >                       </xs:simpleType>
> >               </xs:union>
> >       </xs:simpleType>
> > 
> > 
> > However, I dislike the name of the attribute, elementReference,
> > since the whole point of introducing it was to allow the case where
> > there is no element reference. How about @messageBody or
@bodyContent
> > instead?
> > 
> > Arthur Ryman,
> > Rational Desktop Tools Development
> > 
> > phone: +1-905-413-3077, TL 969-3077
> > assistant: +1-905-413-2411, TL 969-2411
> > fax: +1-905-413-4920, TL 969-4920
> > mobile: +1-416-939-5063
> > intranet: http://w3.torolab.ibm.com/DEAB/
> > 
> > "Sanjiva Weerawarana" <sanjiva@watson.ibm.com>
> > Sent by: www-ws-desc-request@w3.org
> > 
> > 03/11/2004 10:50 PM
> > 
> > 
> > 
> > To
> > 
> > "Jacek Kopecky" <jacek.kopecky@systinet.com>, "Jonathan Marsh"
> > <jmarsh@microsoft.com>
> > 
> > cc
> > 
> > "WS Description List" <www-ws-desc@w3.org>
> > 
> > Subject
> > 
> > Re: Proposed resolutions for issues 146 and 150
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > Looks good to me too .. however I'll let Arthur indicate an IBM
> > position as I can barely spell schiema let alone make value
> > judgements about the goodness of using unions.
> > 
> > Sanjiva.
> > 
> > ----- Original Message -----
> > From: "Jacek Kopecky" <jacek.kopecky@systinet.com>
> > To: "Jonathan Marsh" <jmarsh@microsoft.com>
> > Cc: "WS Description List" <www-ws-desc@w3.org>
> > Sent: Thursday, March 11, 2004 8:58 PM
> > Subject: Re: Proposed resolutions for issues 146 and 150
> > 
> > 
> >  >
> >  > I applaud the elegance of this proposal. 8-)
> >  > I hope it will be accepted.
> >  >
> >  > Jacek
> >  >
> >  > On Wed, 2004-03-10 at 18:55, Jonathan Marsh wrote:
> >  > > Issues 146 [.1] and 150 [.2] were inadvertently left off the
> > FTF agenda.
> >  > > Sorry my bad.  Here's a simple proposal for addressing these
> > issues,
> >  > > assuming we find merit in adding this functionality.
> >  > >
> >  > > Issue 146 Should WSDL be able to describe an operation with
> > *anything*
> >  > > in the message? [.1]
> >  > >
> >  > > Issue 150 Indicating empty bodies [.2]
> >  > >
> >  > > When using XML SchemaS, The element attribute points to a
> > QName of a
> >  > > GED, preventing either empty bodies, or unconstrained
> > content.  Special
> >  > > values of the element attribute could indicate these
> > conditions.
> >  > >
> >  > > Status quo:
> >  > >   <xs:attribute name="element" type="xs:QName"
> > use="optional" />
> >  > >
> >  > > Proposal:
> >  > >   <xs:attribute name="element" type="elementReference"
> > use="optional" />
> >  > >
> >  > >   <xs:simpleType name="elementReference">
> >  > >     <xs:union>
> >  > >       <xs:simpleType memberTypes="xs:QName">
> >  > >         <xs:restriction base="xs:token">
> >  > >           <xs:enumeration value="#any"/>
> >  > >           <xs:enumeration value="#empty"/>
> >  > >         </xs:restriction>
> >  > >       </xs:simpleType>
> >  > >     </xs:union>
> >  > >   </xs:simpleType>
> >  > >
> >  > > (I hope I have got that syntax right.  Should be enough to
> > spark
> >  > > discussion anyway...)
> >  > >
> >  > > [.1] http://www.w3.org/2002/ws/desc/2/06/issues.html#x146
> >  > > [.2] http://www.w3.org/2002/ws/desc/2/06/issues.html#x150
> >  > >
> 
Received on Wednesday, 24 March 2004 17:10:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:30 GMT