W3C home > Mailing lists > Public > www-ws@w3.org > May 2003

Re: Explaining visibility, take 54

From: Geoff Arnold <Geoff.Arnold@Sun.COM>
Date: Fri, 09 May 2003 10:32:25 -0400
To: Mark Baker <distobj@ACM.ORG>
Cc: www-ws@w3.org
Message-id: <0D77246D-822B-11D7-A733-000393C53568@sun.com>


On Friday, May 9, 2003, at 08:25  AM, Mark Baker wrote:

>
> Let's say you have a WSDL document which describes some service, and a
> client and server hardcoded to that WSDL.
>
> Visibility, as an architectural property, refers to the ability of a
> third party component to monitor interactions between other
> components.[1]
>
> So which of these third parties would you say is better able to monitor
> the interactions between the aforementioned client and server?
>
> A.  A generic SOAP/XML intermediary
> B.  An intermediary hardcoded to the WSDL document above
> C.  An intermediary hardcoded to some other WSDL document
>
> I suggest that B has vastly superior visibility to A or C.
>

Visibility may be important for several different reasons. If you
want "semantically deep" visibility into the messages, B will
certainly provide that. On the other hand, if you want to
detect non-compliance and interface evolution, A is probably
superior.

I would also note that a system in which multiple components
are hardcoded to specific metadata (WSDL, schema, whatever)
is intrinsically "brittle". One of the benefits of SOA web services
over (say) CORBA is the opportunity for later binding and
greater robustness in the face of change. (Remember: at large
scale, version skew is a way of life.) IMHO, your strawman
scenario represents poor practice, and should be deprecated rather
than driving WS architecture.....
Received on Friday, 9 May 2003 10:34:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:42 GMT