Re: parameter order

Sanjiva Weerawarana wrote:

>"Umit Yalcinalp" <umit.yalcinalp@oracle.com> writes:
>  
>
>>I point out that you might have  missed the thread if indeed your desire 
>>is to talk about parameterorder.
>>    
>>
>
>No, that's not my desire - I was just pointing out that the 
>parameterOrder solution that you used in WSDL 1.1 (I'm assuming
>you did do that to solve this problem in WSDL 1.1) seems like
>a viable approach now too.
>
>I missed last week's call, but the messages you pointed to are
>not in the space of a parameterOrder attribute. I'd rather not
>have it, but if we're going down that road I'd rather re-use
>the old approach rather than creating something new which
>doesn't even support multi-return type scenarios.
>
My understanding from talking to various people in the wg off line about 
this is that parameter order was not a favorite approach due to its 
perceived complexity. As a result, it was eliminated.

This is why I proposed:

(a)- to use the schema order as an indication of parameter order. I 
deliberately did not propose to bring back the parameterOrder attribute.

(b)- to use the return attribute as the indicator for the return value 
for RPC.

My proposal is very simple and it is scoped well within the context of 
RPC  and can handle most common use cases. It does not bring back 
parameterOrder due to its perceived unpopularity.

As far as the multi value returns are concerned, we need to think about 
the most common cases that we wish to support. The folks in the Lisp 
community can speak about their requirements, but I am in the same boat 
with Tom on this one [1].

>
>Sanjiva.
>
>
>  
>
--umit

[1] http://lists.w3.org/Archives/Public/www-ws-desc/2003Oct/0235.html

-- 
Umit Yalcinalp                                  
Consulting Member of Technical Staff
ORACLE
Phone: +1 650 607 6154                          
Email: umit.yalcinalp@oracle.com