[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [cam] Groups - CAM-v1-draft-013.ZIP uploaded
Monica,
Thanks for taking time to review and send in comments.
My notes inline below,
DW.
=============================================================
Message text written by "Monica J. Martin"
>Dave,
A few questions regarding the document update, by section or concept:
* Section 2.5: Explain what you mean about a test harness.
>> Good point - this is programmer speak - I'll add a clarifying foot note
-
its an outer piece of software - usually a batch script - that calls
the
CAM processor - passes in test input, and logs the output(s) /
results.
<<
* Section 4.1.1: What is a CAM member (is this a registry object
owner, for example - who manages this)?
>> You spotted my task #2 for the 014 revision - use the annotation feature
in schema to document each element / attribute individually.
"CAMmember" here is the URL that points to the location of the
required CAM template document XML.
i.e.
<import
CAMmember="http://lib.UBL.org/src/CAM/OP70-PO-CAM-v1.xml"
CAMalias="UBLOP70"
comment="Standard UBL parts order processing"/>
This could also be a URL lookup to a registry using the RESTful
interface forinstance too. For the 014 revision - we'll add
examples
and more details - being new - I expect this feature to be refined
as
team members send in suggestions / enhancements.
<<
* Section 4.2.1: You have been expanding the CAM efforts/scope and
now are not only talking about content assembly but business
process validation. Can we do a check-balance against the charter
- I see assembly instructions, but not validation.
>> No no! The BPM is a totally separate layer. What we're talking
about here is transaction level validation of the structure contents
only for CAM. Also notice that CAM can be used standalone -
to simply do its thing at the transaction level - no BP involved at
all.
I.e. maybe CAM is validating output web browser form data content.
Clearly assembly level content validation is in scope.
<<
* Section 4.6.1: Will this <choice> construct support some of the
work coming out of other TC that actually define how references to
code lists can be maintained (such as UBL)?
>> That is definately the intention here. We're working with those
folks, and the OASIS registry TC right now on how to represent
codelists so this all works nicely. The CAM has the lookup()
function for using codelists with.
<<
* As I have asked in a previous email., isn't this just a assembly
mechanism, not the validation of the processes it supports (The
document should specify the structural processing instructions,
not the business logic associated with the process validation,
correct)? My question is rooted in how this may affect semantic
interoperability.
>>
CAM is simply a service - call it - get content assembly
stuff done.
Notice how this works. The assembly mechanism contains
context statements relating to the structure and content only.
The BP can pass into the CAM the values of the context for the
structural assembly - so it controls that aspect completely.
The rules in CAM then check the content and the structure.
This is NOTHING to do with a BP engine (like BPEL ) where
the execution instructions rightly go to check whats going on
at the business process level itself.
<<
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]