RE: new editor's draft of WSA available

Actually, SOA seem to be more than an umbrella. It
an ideal that others ascribe to. We can have web services 
that don't meet SOA goals--for example no dynamic discovery
and binding. 

I think we should go ahead an use SOA and use it almost
as a architype of properties that distributed systems
can exhibit. This will help frame the overall evolution
of web services.

daveh

-----Original Message-----
From: Champion, Mike [mailto:Mike.Champion@SoftwareAG-USA.com]
Sent: Friday, October 25, 2002 7:26 AM
To: Mark Baker; Jean-Jacques Moreau
Cc: Christopher B Ferris; www-ws-arch@w3.org
Subject: RE: new editor's draft of WSA available





> -----Original Message-----
> From: Mark Baker [mailto:distobj@acm.org]
> Sent: Friday, October 25, 2002 8:07 AM
> To: Jean-Jacques Moreau
> Cc: Christopher B Ferris; www-ws-arch@w3.org
> Subject: Re: new editor's draft of WSA available
> 
> 

[Jean - Jacques ] 

> > My concern is that the acronym is never formally introduced. 
> > There is no prior indication in the text that "SOA=Service 
> > Oriented Architecture", only in drawings.

Good point.  We need to define it for the glossary and refer to that in the
document.
I think the *sense* of SOA is as an umbrella term that can cover any
technology that consists of service providers, remote  requesters, and a
desscription/discovery mechanism.  COM and CORBA certainly qualify (IDL as
the description mechanism), and I suppose that email or FTP could qualify if
you bent your mind a bit, but that would get into unproductive areas
quickly.  Let's NOT go there now.

Some defintions via Googling:

"A service-oriented architecture has services that developers create in a
service layer. The services that they develop have published interfaces." 
http://softwaredev.earthweb.com/security/article/0,,10527_1041191,00.html

"A service oriented architecture (SOA) combines the ability to invoke remote
objects and functions (called "services") with tools for dynamic service
discovery, placing an emphasis on interoperability".
http://www10.org/cdrom/papers/506/

"SOA takes the existing software components residing on the network and
allows them to be published, invoked and discovered by each other. SOA
allows a software programmer to model programming problems in terms of
services offered by components to anyone, anywhere over the network."
http://archive.devx.com/xml/articles/sm100901/sidebar1.asp


[Mark]
 
> I'd be all for removing any mention of SOAs, if only because I can't
> think of a distributed system which isn't about services.  Email, ftp,
> irc, the Web; all are SOAs.  Besides, we'd have to define it, and we
> know how icky getting concensus on definitions can be. 8-/

We can't avoid this term, it's pervasive.  Icky consensus building is what
we do!
That said, we can't get sidetracked by this. The first page of Google hits
for "service oriented architecture" has the definitions quoted above.  They
are pretty consistent.  Let's let the editors tweak the wording, get it in
the documents, and move on.

Received on Tuesday, 29 October 2002 19:57:47 UTC