Re: Proposal for a protocol binding model


> Why is a separate property necessary? I'd imagine that you'd just

Because both the syntax and semantics for this property have to be
shared between bindings.

> define two bindings; one without many (or any) properties exposed,
> one with a number. You could use different content types for them

Which works for transfer protocols that have a notion of a content
type.  For those that don't, some other syntax will be required.
That's why I suggest one syntax per protocol, rather than per binding.

> (tho I don't really see the utility in this).

It's primarily for security reasons.  A firewall admin should be
able to identify (for blocking, or further filtering) SOAP based
protocols being tunneled over application protocols, while permitting
uses of SOAP that use the application protocols as they were designed
to be used.

FWIW, I believe this is only relevant to application protocol
bindings, not to transport protocol bindings.


Received on Tuesday, 21 August 2001 14:21:39 UTC