Re: MTOM issue: does the HTTP update need to be a feature?

Tony,

Tony Graham wrote:
> Herve Ruellan wrote at 30 Jun 2003 16:56:15 +0200:
>  > 
>  > Having the HTTP implementation be a SOAP feature may be handy the day we 
>  > have several implementation of the abstract feature in the same binding 
>  > to distinguish them.
> 
> I don't understand how there can be several implementations of the
> same abstract feature in the same binding.  Can you explain?

Currently we're targetting at implementing the MTOM abstract feature 
inside HTTP using MIME multipart/related. But we could also implement it 
using MIME multiplexed for example. In this case an HTTP binding 
implementation looking  to be interoperable with as many other HTTP 
binding implementation would include both implementation of the MTOM 
abstract feature. And in such a case, I is handy to be able to 
distinguish the two implementations at a high level.
I agree that this example is somewhat make up, but if we can support it 
easily, let's do it.

Best regards,

Hervé.

Received on Thursday, 10 July 2003 10:30:15 UTC