W3C home > Mailing lists > Public > www-ws-arch@w3.org > December 2002

RE: "Reliable" web services for Next Big Thing?

From: Mathews, Walden <walden.mathews@tfn.com>
Date: Thu, 5 Dec 2002 12:09:35 -0500
Message-ID: <1373D6342FA1D4119A5100E029437F64045EEEA8@clifford.devo.ilx.com>
To: "'Champion, Mike'" <Mike.Champion@SoftwareAG-USA.com>, www-ws-arch@w3.org

> [Wearing my Software AG (a DBMS vendor) hat, not my WSA co-chair hat]
> You may be a RESTifarian without knowing it :-)

If thoughts were hair, I think I'd look about right. 8-))

> I discovered that
> I had RESTifarian leanings for exactly this reason -- it seems so
> much easier to "stick records in a [shared] database" that 
> has transaction
> management, etc. than to worry about all that reliability, routing,
> transactionality,
> coordination, etc. at the messaging layer.

I get your drift, but I have to protest, at least for the case
I'm describing.  Our "database" was a glorified linked list without
any special transactional support or anything else you've assumed.
It's simply that we misinterpreted (transposed) the requirement.
The requirement said we have to store and forward the best available
information in the most timely fashion, and fix the gaps and
errors as opportunity permits.  Pushing that onto the semantics of
a reliable real-time *stream* was the killer mistake.  We thought
we had to "unburden" the repository app by solving the problem
upstream, and it was simply poor analysis (on my part).

Received on Thursday, 5 December 2002 12:10:15 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:41:01 UTC