Re: Where do we find software architecture?

Hi David,

On Thu, Jul 25, 2002 at 12:08:18PM -0700, Burdett, David wrote:
> I agree that sometimes "software happens" and that you can later document
> its architecture. But here are a few questions:
> 1. Would you agree that the "architecture" for web services is not yet fully
> developed?

Yes, certainly.

> 2. Do you agree that this group is tasked with developing an architecture
> for web services?

Yep.

> 3. Developing an architecture (even though it might not be perfect) that
> represents the **full scope** of what you want build is better than having
> an inmcomplete architecture that only represents what currently exists to
> guide the building of the solution?

Well, in the long run, certainly.  But you've got to start somewhere.  I
see our choice as between starting from scratch, or starting from what's
out there today.  I prefer the latter.

> If you agree, then I assert that we would be foolish if we did not look at
> other existing **achitectures** (even though the solutions might not yet be
> built or proven) as well as existing solutions. Otherwise we will lose out
> on the many hours of thought that went into their construction.

Agreed!  That's why I'm spending as much time as I am, supporting
the harvesting effort.  It's critical to study other architectures.
IMO, we should be studying less specs and more systems, but I don't
mind studying both.

Hmm, I wonder where the disconnect was. 8-?

As I understood DaveO's suggestion to do this for SOAP+WSDL, I was
supportive of it.  I would like us to write down the architecture for
current Web services that are out there today, so we're starting our
work from a *real* architecture, warts and all.  This will allow us to
more easily identify areas for improvement, reason about how
extensions can be deployed, figure out how new features (like SOAP 1.2
GET support) fit in, etc..

I get the feeling that DaveO thinks this is some kind of trick on my
part, but it's quite the opposite.  I met with the Director in Boston on
Monday, and he encouraged me to help the WG get something published soon
so that he could give his feedback.  That's what I'm trying to do.  My
concern now, is that we're not going to document the architecture, we're
going to document some kind of "logical model" which won't reflect (by
itself) the sorts of things that he, and many other (but not all 8-) TAG
members, will be looking for.

MB
--
Mark Baker, CTO, Idokorro Mobile (formerly Planetfred)
Ottawa, Ontario, CANADA.               distobj@acm.org
http://www.markbaker.ca        http://www.idokorro.com

Received on Thursday, 25 July 2002 15:41:34 UTC