[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Webber 8/24/2004: Comments on v2.0 Schema from Webber
David, some of my initial feedback on your comments below. Dale and
others, I encourage your comments as well.
==============================================================================================
1) Schema fix ? Variable definition within package should be optional
and unbounded (its OK everywhere else its defined), and Signal should be
unbounded
too (probably better to declare them as global schema types and
reference those to solve this).
mm1: It is optional (note 0->M occurrences). There are no other
constraints applied so further explanation is required.
2) Schema fix ? DocumentSpecificationType should include BPSS, ebMS,
ebContext, CAM, XSLT and XML in token list
mm1: The DocumentSpecificationType maps to a basic enumeration list not
specific to particular implementations of schema or DTD. Perhaps we can
put this to the list to consider specializations of schema or DTD (i.e.
CAM, ebMS, etc).
3) Schema fix ? We have Package within Package. This is very confusing
for implementers. I?d remove this recursion.
mm1: This already exists in the real-world David where there is a
package within a package. I think the for-instance made in the editors'
F2F was there is a process fragment in a package combined with another.
See notes posted in email archives for July 2004.
4) Schema fix ? Do we need the repetition of ? SubstitutionSet,
ExternalRoles and Signal and Variable within Package? Let?s just have
one at the
ProcessSpecification level only.
mm1: Yes, for consistency. Dale, please advise your thoughts.
5) Schema fix ? ExternalRoles / BusinessPartnerRole / Performs has
changed back to fromRole / toRole etc. Was this intended?? I?d like to add
variableGroupIDREF attribute to those attributes.
mm1: Performs / External Roles / BusinessPartnerRole are still present
so I need more information to understand your questions. The to/fromRole
are on the Documentation element (which is consistent with the previous
changes that Martin commented upon). What would this additional
attribute, variableGroupIDREF be used for David?
6) Schema fix ? BinaryCollaboration - TimeToPerform ? should be optional!
mm1: David, this has not changed in the current specification, TTP
already existed prior to this time associated with a BinaryCollaboration
(and was not optional).
What reasons should this be optional? In addition, the TTP is used to
designate that a given request and response are received in the
agreed-upon time.
7) Schema fix - BusinessTransactionActivity ? TimeToPerform ? should be
optional!
mm1: Same as above.
8) Schema fix ? rename TimeToPerform to timeToComplete ? this avoids
confusion with the messaging attribute purpose.
mm1: Why would we change this David. The ebMS specification uses
TimeToLive and that is well understood as it TimeToPerform.
9) Schema fix ? replace the bare Variable definition with this structure
? that provides far more fine grained control ( also note ?
ConditionExpression should
have unbounded occurs )
<ContextVariables>
<variableGroup name-?default? nameIDREF=?default01?> <!-- the first
variableGroup instance is automatically the default group that applies
to all
roles and context.-->
<Variable name=?myvar1? nameID=?myvar1? label=?Value 1:?
default=?01?>
<ConditionExpression expressionLanguage=?XSD?
expression=?string?/>
<ConditionExpression expressionLanguage=?XSLT?
expression=?string?/>
</Variable>
<Variable name=?myvar2? nameID=?myvar2? label=?Value 2:?
default=?002?>
<ConditionExpression expressionLanguage=?XSD?
expression=?string?/>
<ConditionExpression expressionLanguage=?XSLT?
expression=?string?/>
</Variable>
</variableGroup>
<variableGroup name-?for-GM-only? nameIDREF=?gm01?> <!-- subsequent
variableGroup are referenced by name from ExternalRoles for specific
participants -->
<Variable name=?GMvar1? nameID=?myvar1? label=?Value GM1:?
default=?NORMAL?>
<ConditionExpression expressionLanguage=?XSLT? expression=?
doc:\\CustomerPO\Header\orderType='URGENT? documentRef=?purchaseOrder?/>
</Variable>
</variableGroup>
</ContextVariables>
mm1: The variable element definition is not complete. Why would we embed
context here David? I'd suggest John Yunker and Dale Moberg review the
more detailed changes. I was of the understanding that some of the more
extended context possibilities were being held until v3.0.
10) Schema fix - Documentation ? in addition to URI this should have an
xml:lang attribute.
mm1: Please explain why a URI is insufficient. Dale, please also advise.
Then these changes are clean-up / clarification related:
11) SubsitutionSet ? Deprecated - Remove
mm1: Substitution sets have not been deprecated.
12) isConcurrent? Does this need to be there now ? just use Join? Remove?
mm1: There has not been discussion to remove. There was some discussion
on how to effectively use isConcurrent in specific cases (see editors'
notes from F2F).
13) isInnerCollaboration? Same thing ? surely endsWhen / beginsWhen
handles by inference this?
mm1: No, see editors' F2F notes. We actually placed some well-formedness
rules around its usage.
14) What does Pattern do? It contains URI - linking semantics to BPSS
models?
What syntax is pattern in? UML, PDF, Word? If so ? maybe we need
patternType
attribute too?
mm1: We discussed whether or not we leave the pattern attribute for
legacy enablement. See:
http://www.oasis-open.org/archives/ebxml-bp/200407/msg00131.html. I
believe this was a name attribute to the actual pattern. Looking at the
technical specification (hint David), the pattern maps back to the
guiding reference for the pattern definition (for the pattern attribute).
15) isLegallyBinding ? rename? How about ? stepHasLegalMeaning
mm1: We discussed this yesterday with HasLegalIntent. Hopefully we can
get the final set of details from interested experts to finalize a decision.
16) Do we really need Include? This can be done with an XSD include import,
right? Deprecate / Remove?
mm1: See changes from editors' F2F and notes related to Include. See:
http://www.oasis-open.org/archives/ebxml-bp/200407/msg00035.html
17) Fork, Decision and Transition ? these are now really all the same
thing!!!
I think we can remove Transition and Decision.
mm1: I don't believe they are the same David. Dale, can you also expand.
I believe we discussed that some of these business states could be
combined but not deleted.
18) Fork has timeToPerform attribute. Remove?! [ the timeToComplete of
the preceding steps really control this]
mm1: The TTP on a fork is optional (and has historically been available
in ebBP) [see page 44 in v1.1].
19) Join should have an overall optional timeToComplete on it!!! Add.
mm1: See comment in above.
20) OperationMapping ? what is this for? Should it be using nameIDREF
instead
of name?
mm1: This is the enabler for use of web services David and is only
mapping to an abstract name (operation).
Thanks.
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]