RE: Yet another attempt to fix D-AC004

Hi Mark,

Issue 1.

>Ah, I think we have a terminology issue here.

>I've been using the term "component" in the context of an architecture.
>From our glossary;

>Architecture

>  The software architecture of a program or computing system is the
>  structure or structures of the system, which comprise software
>  components, the externally visible properties of those components,
>  and the relationships among them."

>I'm talking about that kind of component; like a Web client, proxy,
>gateway, server, tunnel, etc..

My reading of the above glossary definition of architecture does not 
technically limit it to the list you suggest, although I agree that it 
is implied. Specifically, it can be read that components are not necessarily
externally visible. 

To me, it is important to be able to discuss all the functional entities 
of the WS architecture, externally visible (whatever that really means) or
not, as components. Hence security, privacy, reliability, etc. must be described 
in terms of components and relationships. 

Hence, loose-coupling applies to this broader description of components 
whereas early/late-binding is a subset in the domain of the service discovery 
components of the WS architecture. 
 

Issue 2.

>> AR00X.6 support both early and late client binding to web services.

>Sure.  This should probably be related to D-AR003.6 in some manner,
>since it not only suggests "support", it says that we should define
>how late binding occurs.

>http://www.w3.org/2002/ws/arch/2/06/wd-wsa-reqs-20020605.html#ar003.6

>Not sure how best to do that though.

We could move D-AR003.6 to this CSF. I think this fits better here than
under 'extensibility'. What do you think?


Issue 3.

>> AR00X.7 components must described by their functional roles and responsibilities.
>Modulo a "be" in there, that's great.

Thanks!


Issue 4.

>> If you agree that describing the relationship between architectural components
>> is in scope, do you have a better suggestion.

>How about just "The relationships between components must be well described"?

I was hoping to get alittle deeper than that. Maybe we can discuss this on the
telecon tomorrow.

Received on Wednesday, 26 June 2002 15:50:11 UTC