Re: Language for resolving issue 440

On Jan 15, 2004, at 8:40 PM, [email protected] wrote:

> Marc Hadley asks:
>
>>> I don't recall - do we already have text that
>>> captures the other part of the resolution,
>>> namely that MTOM doesn't preclude additional
>>> parts in the package not reference via
>>> miffy:Include ?
>
> Did we conclude that?

Yes, we did.

>   More specifically, did we conclude that for Miffy
> or for the HTTP binding as well?

For both. The variability we introduced between XOP and MTOM was 
related to the cardinality of inclusions: we concluded that in XOP 
there could be multiple inclusions of the same binary part but that the 
HTTP binding would restrict it to a single inclusion of any binary 
part. Both HTTP binding/MTOM and XOP would allow unreferenced 
attachments but they are outside the SOAP processing model.

Marc.

>  Not surprisingly, I'm fairly strongly
> opposed to allowing variability in the content sent by the HTTP 
> binding.
> That said, I may just be repressing memories of a decision that went
> counter to my preferences.
>
> I've just read the review copies of the specs, and regardless of what 
> our
> issue resolutions say (and they should be clear), the current miffy 
> text
> does allow for separate parts.  The HTTP binding could probably be read
> either way, since it says to make a part for each optimized piece, but
> doesn't really say whether that means >only< for each optimized piece. 
>  I
> think we should remind ourselves what the resolution is on the http
> binding, and clarify the MTOM document either way.  I think I can live
> with the variability in Miffy/XOP.
>
> --------------------------------------
> Noah Mendelsohn
> IBM Corporation
> One Rogers Street
> Cambridge, MA 02142
> 1-617-693-4036
> --------------------------------------
>
>
>
>
>
---
Marc Hadley <marc.hadley at sun.com>
Web Products, Technologies and Standards, Sun Microsystems.