Re: Requirements: Decision Points Requirement Proposals

In my personal opinion adding XPath support to the spec would take 
considerably less time & effort than we've spent so far discussing the 
point of whether or not to add XPath support to the spec.

So I would like to make a friendly suggestion.

I am personally of the opinion that support for various semantics 
including the ability to describe rules that relate to the message 
content is damn useful. And the language must support that requirement. 
But my definition of support at this point in time does not mean it 
needs to incorporate XPath expressions. The ability to extend the 
definition with XPath expressions - either through some form of 
extension mechanism or something that would be formalized in a future 
version of the spec - is a minimal form of support that may be adequate 
for version 1.0.

However, we need to recognize that in doing so we are supporting a 
requirement, even if we do so in a rudimentary manner. So we need to 
acknowledge the requirement, and then perhaps reach a group consensus 
that the priority is sufficiently low, that such a minimal form of 
support would be adequate for a v 1.0 deliverable.

Alternatively, we may decide to allocate a well constrained amount of 
time & resources to building such support in the language. Whatever we 
can accomplish in a limited amount of time would end up in the spec. If 
we can do more, great. If we can do little, so be it. But I think doing 
something even if minimal would be a better use of my time than an 
endless discussion of whether or not to do anything or everything.

arkin

Yaron Y. Goland wrote:

> If we are to harbor any hope of shipping any time soon I would suggest 
> we follow the rule that 'the spec is ready to ship when there is 
> nothing left to cut'. We can clearly cut XPATH and still have an 
> enormously useful spec. We can also layer XPATH support in on a future 
> release so our choice does not permanently cut off us. Therefore I 
> would propose leaving off XPATH.
>  
>
>     -----Original Message-----
>     *From:* Cummins, Fred A [mailto:fred.cummins@eds.com]
>     *Sent:* Thursday, July 03, 2003 1:56 PM
>     *To:* Yaron Y. Goland; Jean-Jacques Dubray; 'WS Chor Public'
>     *Subject:* RE: Requirements: Decision Points Requirement Proposals
>
>     Yaron, JJ,
>      
>     Regardless of the maturity of the effort, if I understand the
>     XPath approach, this sounds like the choreography would be defined
>     in terms of the structure of the messages.  I think the
>     choreography should be oblivious to the structure of the messages,
>     which might be XML or something else.  The determination of what
>     logical document is received can be delegated to the application
>     that will process it.
>      
>     Fred
>

Received on Sunday, 13 July 2003 14:31:02 UTC