- From: Mark Nottingham <mnot@mnot.net>
- Date: Fri, 6 Jul 2001 11:49:58 -0700
- To: Henrik Frystyk Nielsen <henrikn@microsoft.com>
- Cc: "Williams, Stuart" <skw@hplb.hpl.hp.com>, xml-dist-app@w3.org
> >So is SOAP really a protocol or is it yet another > >packaging/encapsulation format? > > ...as opposed to yet another protocol ;? Protocols are used for all kind > of purposes and exhibit all kind of properties and diving into what > constitutes a protocol and how they can be classified is nothing less > than a PhD subject. > > The discussion that we have had about the SOAP processing model will > tell you that there is more than merely a wrapper - there is in fact a > processing model that defines how to deal properly with SOAP messages > and when to generate faults. > > It is true that SOAP by itself doesn't define many of the application > layer characteristics that most other application protocols exhibit such > as message exchange patterns, routing, correlation, etc. The reason > being that I think we believe we have a good extensibility mechanism > that allows us to add such features. Time will tell whether this is a > valid assumption. Well said... I've been wrestling with this for a while, and the closest I've come is 'protocol framework' or 'messaging format with protocol behaviour hooks'. In the dictionary sense, SOAP is a protocol, but that by itself isn't too helpful. The common-use meaning of 'protocol' is equally unhelpful, as people use it to mean different things. It's also confusing because SOAP is made up of: * an RPC convention (which is very protocol-ish) * a serialization mechanism * an envelope * bindings (possibly to protocol-y things) So, the 'edges' of SOAP exhibit protocol-like attributes, but the 'core' doesn't. -- Mark Nottingham http://www.mnot.net/
Received on Friday, 6 July 2001 14:50:01 UTC