- From: Mark Baker <distobj@acm.org>
- Date: Thu, 19 Jun 2003 21:09:27 -0400
- To: Savas Parastatidis <Savas.Parastatidis@newcastle.ac.uk>
- Cc: www-ws@w3.org, umit.yalcinalp@oracle.com, Jim Webber <jim.webber@arjuna.com>, Steve Graham <sggraham@us.ibm.com>, Krishna Sankar <ksankar@cisco.com>
Hi Savas, On Thu, Jun 19, 2003 at 09:08:34PM +0100, Savas Parastatidis wrote: > Mark, > > > > b) is less obvious, but most Web services interfaces I've seen are > > stateful. I believe this is attributable to the simple fact that it > > takes effort to design stateless interfaces, whereas stateful > interfaces > > are the default. > > > > I would argue that SOA says nothing about the statefulness of the > services exposed through an interface. I agree. I was just referring to interfaces I've seen; most of them (other than the trivial getFoo() ones) are stateful. > I agree that most of the Web > services have to maintain some state behind the scenes in order to be of > any use but the semantics of SOA do not mandate that (at least that's my > understanding). Erm, well you seem to be talking about a) state there. I'm talking about b) state, aka "application state" or "conversation state". In that context, "stateful" is where the interpretation of message semantics depends upon information not available within the message itself (sometimes called "shared context"). Mark. -- Mark Baker. Ottawa, Ontario, CANADA. http://www.markbaker.ca
Received on Thursday, 19 June 2003 21:04:57 UTC