[
Lists Home |
Date Index |
Thread Index
]
SOAP is an interface: that is, the front end is disclosed, public, intended to be
programmed to. It is the back end on each processing node where 'confidentiality,
reliability, and protocol neutrality' are offered by 'securing' confidential
information and mission-critical services.
By contrast, REST is the epitome of the document orientation that is the first
premise both of markup and of the http transport. RESTful documents (high, low, or
in-between--they are documents) are *published* as the back-end output product of
each processing node. Those documents are presented in the clear--are, in fact,
delivered in an appropriate representation for each known sort of requester. What
is not known is what each requester will do with the RESTful documents which it
fetches. It is the front end of the requester which is opaque: that front end
takes known document types as published and then instantiates them into non-public
datastructures for its own purposes in processing. In a fully RESTful system the
output of such processing will itself then be published in a suitable location
(and perhaps with suitable security governing access) for further uses by
additional interested parties who need not even be known to the publisher of those
documents.
Respectfully,
Walter Perry
Michael Champion wrote:
> On Wed, 30 Mar 2005 10:01:42 -0800, Dare Obasanjo <dareo@microsoft.com> wrote:
>
> > > I *think* the whole SOAP vs REST permathread comes down to
> > > Resource-orientation vs Service-orientation ...
> >
> > I completely disagree with that. SOAP vs. REST is rich vs. reach.
>
> Fair enough. I was talking about the permathread abstract discussions
> of architectural styles, service architectures, etc., i.e. "high REST"
> in Nelson Minar's terminology. I agree that the question of "low
> REST" vs WS-* does come down to rich vs reach.
|
