RE: Visibility (was Re: Introducing the Service Oriented Architec tural style, and it's constraints and properties.

"Cutler, Roger (RogerCutler)" <[email protected]> Fri, 28 Feb 2003 10:36:40 -0600
Newsgroups gmane.comp.web.services.ws-arch
Message-ID <7FCB5A9F010AAE419A79A54B44F3718E01624B00@bocnte2k3.boc.chevrontexaco.net>
IMHO he has indeed answered you, he just hasn't agreed with you.

-----Original Message-----
From: Mark Baker [mailto:[email protected]] 
Sent: Friday, February 28, 2003 1:09 AM
To: David Orchard
Cc: [email protected]
Subject: Re: Visibility (was Re: Introducing the Service Oriented
Architec tural style, and it's constraints and properties.



On Thu, Feb 27, 2003 at 10:36:41PM -0800, David Orchard wrote:
> And now I'll answer the first paragraph.  Visibility is a degree of 
> visibility, not an absolute yes/no.

Of course.

>  Firewalls will look at many things in
> messages, like ip addresses, http methods, URIs, port #s, etc.

Sure.

> Even if the method name goes in the SOAP envelope, it's still visible 
> to the intermediary.

No, it isn't.  It isn't enough that the string "FOO" can be run through
some parser, because anything can be run through a parser.  The issue
is, does the app have prior knowledge of that what that string means?

> It may be harder than if the method wasn't.  I think you are 
> purposefully avoiding the simplicity argument that goes along with 
> multiple protocols.  There is a trade-off in properties at play.  
> Roughly it's simplicity vs visibility and performance.

I'm not saying that other properties weren't improved upon - perhaps
they were, in spades, I don't know.  I'm just asking about visibility;
is there less visibility with the SOA style than with the REST style?

You're not going to answer that, are you? 8-/

MB
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
Web architecture consulting, technical reports, evaluation & analysis