Rich Salz on WSDL 2.0

Don Box's Spoutlet

Syndication

(via Gudge) Rich has an interesting rant on WSDL 2.0 over on XML.COM.
 
While we're dog-piling on WSDL/2.0, my personal pet peeve is that the WG had the opportunity to make stronger commitments to the SOAP data model and therefore make the aggregate language much simpler. Granted, WSDL/2.0 is better in this regard than WSDL/1.1, but the fact that I can't express application headers in an interface is disappointing.

Posted Nov 21 2004, 04:18 PM by don-box

Comments

Mark A. Richman wrote re: Rich Salz on WSDL 2.0
on 11-21-2004 4:35 PM
Don,

What do you mean by "application headers" ?

- Mark
Don Box wrote re: Rich Salz on WSDL 2.0
on 11-21-2004 6:48 PM
Mark,

I mean headers that are part of the app contract and not added by infrastructure.

The motivations for wanting app headers is that most if not all SOAP stacks will have read and buffered the header blocks as part of dispatch. It's often the case that the body is streamed. That means that if you want information to be visible "early" in the processing chain, you want it in the S:Header element.

DB


Ed Pinto wrote re: Rich Salz on WSDL 2.0
on 11-21-2004 10:00 PM
Don,
What bugs me most is the unnecessary overlap with WS-Policy. As far as I can tell, the feature and property components aren't needed if WS-Policy is used - What's your take?

Cheers,
Ed

Don Box wrote re: Rich Salz on WSDL 2.0
on 11-22-2004 8:28 AM
Ed,

I agree that the F&P stuff has significant overlap with WS-Policy.

Note that MSFT, IBM and others filed a minority opinion that the F&P stuff shouldn't be there.

In general, I haven't seen any vendor implement or even embrace F&P. My working assumption is that Darwin will kick in (as always)...

DB
Dave Orchard wrote re: Rich Salz on WSDL 2.0
on 11-22-2004 8:30 AM
Don,

I hear you. BEA had to fight soo hard to get the "Application Data" feature in to WSDL 2.0, which is exactly what you are talking about. The problem was folks who believe that soap headers are basically never seen by the app and everything should go in the body, despite the existance of the soap processing model for headers. AD is in the part 2 optional section and it uses f&p, which I'm convinced means it won't get widely implemented. :-(
Don Box wrote re: Rich Salz on WSDL 2.0
on 11-22-2004 8:36 AM
Dave,

Sad, sad, sad...

If only you and I were King for a day, what beautiful damage we could do...

DB
Mark A. Richman wrote re: Rich Salz on WSDL 2.0
on 11-22-2004 12:19 PM
It's funny...I co-authored a SOAP book just a couple of years ago. Then I stopped doing WS work for a little while, came back to it, and I feel like a newbie all over again!!! :(

http://www.markrichman.com/publications.html#book1

- Mark

Add a Comment

(required)  
(optional)
(required)  
Remember Me?