Notes on WSDL support, conformance statements, and miscellaneous items

This email is a hodgepodge of different items related to WSDL, mostly
for discussion.  If the group agrees that these items are issues, I'll
raise those separately.  There's enough going on here, that I figured a
more general discussion was worthwhile.

Specification:
==============

Item: Section 3.4.5 references the URI specification. This is mostly odd
in that it is not the first reference to URI.  Suggest that we need an
editorial review to look for all references, and that the first
reference to a concept is the one that gets linked.  I looked at
http://www.w3.org/TR/2009/PR-owl2-syntax-20090922/#Normative_References
for comparison.  Do we want a top-to-bottom review of this?

Item: Once restated as suggested by ISSUE-15, normative statement 3001
no longer needs to be normative.  Rather, 3002 has it covered.  Shall we
collapse?

Item: Do we allow multiple instances of elements in WSDL?  Seems like we
shouldn't allow two different <soapjms:timeToLive/> elements as
siblings.  Do we need a normative statement for this?

Item: Normative statement 3003 appears to be somewhat redundant with
3002.  At least, it is also about which items take precedence.  Should
we combine this statement with 3002?

Item: Normative statement 3004 doesn't have any RFC 2119 keywords.  Any
suggestions on how to fix?

Testing:
========
For WSDL test cases, I think we clearly need one or two test cases to
verify setting of WSDL properties at different levels, and having the
more detailed items trumping the more general items.

Currently, our normative statements don't actually seem to require more
than this.  What are we missing?

-Eric.

Received on Tuesday, 20 October 2009 05:13:28 UTC