Re: Survey on usefulness of p2 family patterns

At 08:02 PM 6/24/2003 -0400, you wrote:


>----------------------- Begin Survey ------------------------
>
><p2-survey>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2' 
>
> >
><myOpinion>

This is Gudge's "Interesting Pattern", which (I believe) he likes because 
it allows the service side to treat applications needing p2d and p2e the 
same way.  However, for client-side applications, I don't think it conveys 
enough information in the common request-response interaction.

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2a' 
>
> >
><myOpinion>

Useless, since the fault doesn't necessarily go back to A.

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2a1' 
>
> >
><myOpinion>

Potentially useful.  However, from client A's perspective, all A did was 
send a message and potentially receive a fault back (which is pattern 
p1b).  And from any Bi's perspective, all it did was receive a message 
(which is pattern p5a).  So I'm not sure we need this one.

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2a2' 
>
> >
><myOpinion>

I think it would be more helpful to have the fault sent back to A, so I 
don't think this pattern is very useful.

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2b' 
>
> >
><myOpinion>

Same opinion as for p2a1: Potentially useful.  However, from client A's 
perspective, all A sees are the messages of a classic request-response 
pattern, which is p2e.  Therefore,I'm not sure we need this one.

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2c' 
>
> >
><myOpinion>

Useless, since the fault is not going to A or B.

></myOpinion>
></pattern>

><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2d' 
>
> >
><myOpinion>

Useful.  This is a request-response-to-third-party pattern.  However, from 
A's perspective this is just pattern p1b, and from B's perspective it is 
just p5a, so I'm not sure this pattern is needed.

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2d1' 
>
> >
><myOpinion>

Less useful than p2c, because the fault goes to B.

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2e' 
>
> >
><myOpinion>

Definitely useful.  Classic request-response.

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p2umci_7' 
>
> >
><myOpinion>

Probably not useful.  (Who cares whether other services also receive a message?

></myOpinion>
></pattern>
>
><pattern 
>name='http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/meps-vs-iops_clean.htm#p3' 
>
> >
><myOpinion>

This requires communication to be "on the same channel".  I don't think the 
"same channel" requirement is what we need.  (Who cares if the 
communication is on the same channel?  The important point is that the 
messages got where they were supposed to get.

></myOpinion>
></pattern>
>
></p2-survey>
>
>
>
>--
>David Booth
>W3C Fellow / Hewlett-Packard
>Telephone: +1.617.253.1273

-- 
David Booth
W3C Fellow / Hewlett-Packard
Telephone: +1.617.253.1273

Received on Monday, 30 June 2003 04:01:34 UTC