RE: [RTF] AC019 proposal to WSA WG

> -----Original Message-----
> From: Mark Baker [mailto:distobj@acm.org]
> Sent: Wednesday, July 10, 2002 10:35 AM
> To: Champion, Mike
> Cc: www-ws-arch@w3.org
> Subject: Re: [RTF] AC019 proposal to WSA WG
> 
> that there's a cost in the brittleness of any system built to an
> architecture that implements reliable messaging, due to the reasons
> discussed in the Waldo paper.

Hmmm ... that is a useful paper and I agree with much/most of it.
Still, to me there is a lot of distance between ignoring / hiding
the fact that "objects" are distributed and asking for the messaging
infrastructure to handle the grunt work of making sure that messages
are delivered.  Of course, the latter might allow naive developers to
assume that the Internet is just a giant LAN and ignore the intrinsic
latencies, opportunities for failure, etc. of a global network, thus 
producing brittle applications.  I think it's our job to empower
competent developers, not to prevent novices from hurting themselves.

> I'm just saying that there are ways of addressing it 
> that don't invole requiring that every message arrive at its 
> destination, and that it is primarily a function of the architectural
style in use 
> as to which solution is the most appropriate.

I'm saying that there is, IMHO, no incompatibility between the REST
architectural style and a reliable messaging infrastructure.  Of course,
REST provides a discipline for dealing with messaging failure that does not
demand as much underlying reliability as synchronous RPC, for example.  That
makes the REST style a good choice for, say, wireless web services, and
nothing we are  discussing with respect to the WSA will change that.  (I
suppose we would give naive wireless developers a few strands of rope that
they could hang themselves with ... but again I don't think that's our
problem). 

I very much want the WSA to accomodate developers who choose to use the REST
architectural style, but we should not cripple the WSA so that it can ONLY
support REST. Even the synchronous RPC "architectural style" has its place
in the world ... maybe not over the public internet, and probably not over
the wireless web anytime soon, but certainly in environments that the
members care about and where standardization would be useful.  Our job is to
identify and clarify the architectural bits that can be widely and commonly
used, and I see reliable messaging as one of them, hence it belongs in our
requirements.

Received on Wednesday, 10 July 2002 10:53:01 UTC