axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Graham" <>
Subject Re: [Vote] Rick Rineholt becoming a committer
Date Wed, 26 Sep 2001 11:12:46 GMT

Steve Graham
(919)254-0615 (T/L 444)
Emerging Technologies

Doug Davis/Raleigh/IBM@IBMUS on 09/25/2001 09:19:18 AM

Please respond to

Subject:  [Vote] Rick Rineholt becoming a committer

+1  for Rick becoming a committer.

Russell Butek/Austin/IBM@IBMUS on 09/25/2001 08:54:19 AM

Please respond to

Subject:  WSTK model for WSDL emitter

IBM's WSTK folks - notably Rick Rineholt - have a SOAP Encoded Object Model
(SEOM) that they use in their WSDL emitter.  They are offerring their model
to AXIS.  I spent the last couple days rewriting the Wsdl2java emitter to
use their model instead of WSDL4J and DOM directly.  Here is the result.
This has all of our stuff plus all of the SEOM code.  (See attached file:

Benefits of their model:
- It handles most schema types, ours barely does.
- We'll probably find we need some sort of model that's easier to code to
than DOM and WSDL4J (we've already created the Parameters class).  SEOM is
such a model.
- They are willing to support their model, growing it to match the needs of
AXIS - in other words, Rick could become a committer.

Possible drawbacks of their model:
- SEOM barely exposes WSDL4J or DOM.  This was intentional in their design,
but our emitter may need to be exposed to them to a greater degree.  For
example, the WSDL service tag isn't reflected in SEOM.  I left the
generation of deploy.xml, undeploy.xml, pretty much alone
just because I couldn't generate these using SEOM.
- First we had DOM, then WSDL4J, now SEOM.  Three layers of a symbol table.
SEOM brings in a lot of new classes.  Does AXIS really need a layer as
thick as SEOM?

Russell Butek

View raw message