RE: Application or Infrastructure? (was FW: LC Comments: Web Method Feature) (fwd)

Caught by spam filter.

-- 
Yves Lafon - W3C
"Baroula que barouleras, au tiéu toujou t'entourneras."

---------- Forwarded message ----------
Date: Tue, 9 Jul 2002 09:49:43 -0400 (EDT)
From: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>
To: "'xml-dist-app@w3.org'" <xml-dist-app@w3.org>

> -----Original Message-----
> From: Paul Prescod [mailto:paul@prescod.net]
> Sent: Tuesday, July 09, 2002 9:23 AM
> To: Champion, Mike; 'xml-dist-app@w3.org'
> Subject: Re: Application or Infrastructure? (was FW: LC Comments: Web
> Method F eature)
>

> Wonderful. I'm glad that they have a solution to their problem. But I
> don't think it has anything to do with xml-dist-app@w3.org.

Good point.  Still, there's something to be said for standardization to
promote reusability of software and knowledge even when interoperability
is not an issue.  Perhaps that's not the mission of the W3C ... I'd be
happy to accept that, but I think the W3C membership has already taken a
position that SOAP / WSDL standardization and "transport" neutrality
is a Good Thing.

> Retries are by definition handled by software. Whether they
> are handled by *application* software depends upon how your software
> toolkit divides the world into application and toolkit.

OK, I continue my confusion about "application" protocols vs application
software.  I'm not seeing the architectural problem with handling
reliability, security, etc. below the application protocol layer.
You're saying that it's OK to put it below the applicaiton layer of the
software, but it needs to be done with an application-level protocol?
I hope you can see how this makes my head hurt :~)

> I do not see anything in SOAP that automates this to any
> greater extent than HTTP.

Fair enough; this discussion really belongs on the WS Architecture
list where we think about how to add such features.  The obvious way
(once SOAP-think has infested your neurons) is with SOAP headers to
either request a reliable infrastructure protocol or to implement
one on top of a less reliable "transport".

> Furthermore, if we standardize the HTTP header then it can be used on
> the ordinary web as well as the "services" web.

SOAP is transport-neutral, for better or worse,
and it's not clear to me how that suggestion helps us.  If the suggestion
is aimed at *users* as a Best Practice guideline, I completely agree.

Received on Tuesday, 9 July 2002 12:13:24 UTC