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

RE: features & properties: anywhere or only selected places??

From: Liu, Kevin <kevin.liu@sap.com>
Date: Fri, 12 Mar 2004 20:46:13 +0100
Message-ID: <99CA63DD941EDC4EBA897048D9B0061DA95ED5@uspalx20a.pal.sap.corp>
To: "'Roberto Chinnici'" <Roberto.Chinnici@Sun.COM>
Cc: "'Anish Karmarkar'" <Anish.Karmarkar@oracle.com>, "'Umit Yalcinalp'" <umit.yalcinalp@oracle.com>, Sanjiva Weerawarana <sanjiva@watson.ibm.com>, www-ws-desc@w3.org, Martin Gudgin <mgudgin@microsoft.com>, "David Orchard (david.orchard@bea.com)" <david.orchard@bea.com>

Hi Roberto,

I agree with the list of common issues you summarized below. Actually, I can even add more to that, such as inheritance rules, union and intersection rules, and so on. I definitely agree that a common framework to address these issues would be beneficial for all of us.

As I already pointed out, my concerns are:

1. I don't believe a complete solution for this common framework can be developed in a few weeks.  From what we learned from relevant work, it might need some 12 months to work out a solution for describing configuration information at each level of the WSDL element hierarchy, to define inheritance rules, a processing model, union and intersection rules and so on. 

2. At this late stage, it doesn't worth anymore to dilute the already limited resource we have in the WG to tackle a complicate new feature for WSDL while we are juggling with basic issues to get the last call ready. We realized the value of a common framework for F&P and supported the creation of the WSDL F&P task force last year to explore that space. Unfortunately, the TF has been dormant for almost all the time. I don't think anybody should be blamed for that, it's just a clear indication that a complete solution is not available yet. The compositor proposal is a step forward, but even if it's accepted, it still only covers part of the issues.  I believe many member of the group have already expressed their opinion via last week's vote.

3. Without the compositor (and maybe a lot more need to be added to make a practical solution), is the current F&P constructs in WSDL really useful? I strongly doubt it. I am not sure how many companies will really use the WSDL P&F we baked here when a more complete solution is available. For one, we will not.
In summary, I agree with you that we definitely need a common framework as you suggested, but that framework deserves its own working group which will be dedicated to develop a complete solution that is composeable with other protocols.   Just squeeze a half baked solution to WSDL will not help anybody, in stead, it only adds unnecessary complexity in WSDL and create unnecessary confusion to the Web services community. 

Best Regards,

-----Original Message-----
From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org] On Behalf Of Roberto Chinnici
Sent: Thursday, Mar 11, 2004 05:16 PM
To: Liu, Kevin
Cc: 'Anish Karmarkar'; 'Umit Yalcinalp'; Sanjiva Weerawarana; www-ws-desc@w3.org
Subject: Re: features & properties: anywhere or only selected places??


I agree that "WSDL's scope is to describe a single service and that we
should focus on what's in scope and provide a SIMPLE BUT SUFFICIENT
service description language".

But I'd like the language to be sufficient not for the needs of two
years ago, but for the needs of the foreseeable future. Those needs are
represented by the different working groups that are basing their work
on WSDL -- things like security, reliability, management.

In my opinion, it's fully within the mission of this working group to
define a general *framework* that other groups can build on to enrich
a service description with information about security, reliability, etc.

Clearly, all those groups will run into a set of common issues, e.g.
   - defining ways to annotate service descriptions with additional
   - defining a processing model for these "annotations" and setting
     limits to what the specification for an annotation can mandate;
   - clarifying how annotations will interact with each other when
     they are applied to the same WSDL component;
   - defining how annotations will (formally or informally) compose with
     each other when applied at different levels in a chain of components,
     i.e. along the service->endpoint->binding->interface->operation axis;
   - defining how annotations interplay with operation inheritance;
   - specifying the means by which WSDL authors can compose annotations
     into more complex expressions.

Does duplication of work across multiple groups and potential gratuitous
differences in the adopted solutions appeal to anybody?

Do implementors wish to juggle the potentially conflicting requirements
on processors placed by a dozen different specs? I don't.

Why then shouldn't our group address some of these issues and provide
a coherent answer in the form of a framework such as f&p (with compositors)?


Liu, Kevin wrote:
> It's really not important what's the other spec. My point is that F&P deserves its own complete solution which WSDL should be able to compose with. 
> WSDL's scope is to describe a single service. We should focus on what' s in scope and provide a SIMPLE BUT SUFFICIENT service description language. Anything else, including policy, security, reliability... should be left for other specs. 
> Best Regards,
> Kevin
> -----Original Message-----
> From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org] On Behalf Of Anish Karmarkar
> Sent: Thursday, Mar 11, 2004 10:50 AM
> To: Liu, Kevin
> Cc: 'Umit Yalcinalp'; Sanjiva Weerawarana; www-ws-desc@w3.org
> Subject: Re: features & properties: anywhere or only selected places??
> Liu, Kevin wrote:
>>I don't recall the group has ever agreed that F&P can appear under <service>. What's the use case for that?
>>It's only a few days to our last call, how many of us in the group are confident that he/she fully understand how the WSDL F&P things work. Am I the only one worrying? Since Feature and property are such an important area, I am very concerned that we are making WSDL unnecessarily complicated while only provide a partial solution. 
>>I am against to expand it to any more element. Instead, I suggest we remove F&P from WSDL2.0, and leave it to other specs where it belongs to.
> I curious about what you mean by 'other specs'? Can you please 
> elaborate? I cannot think of any other non-proprietary spec that deals 
> with this issue.
> -Anish
> --
>>Best Regards,
>>-----Original Message-----
>>From: www-ws-desc-request@w3.org [mailto:www-ws-desc-request@w3.org] On Behalf Of Umit Yalcinalp
>>Sent: Wednesday, Mar 10, 2004 11:52 AM
>>To: Sanjiva Weerawarana
>>Cc: www-ws-desc@w3.org
>>Subject: Re: features & properties: anywhere or only selected places??
>>Sanjiva Weerawarana wrote:
>>>Hi Guys,
>>>I was always under the impression that <feature> and <property>
>>>elements are allowed under any element .. however, the spec
>>>says otherwise. Did I just get confused?
>>I guess you mean any component? We definitely need F&P under the service 
>>element (more precisely ServiceType ;-)) in addition to 
>>interface/operation/binding/binding operation components.
Received on Friday, 12 March 2004 14:46:54 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:54:47 UTC