Ashok,
Replies inline marked as <mje>...</mje>
Yours, Mike.
Strategist - Emerging Technologies, SCA & SDO.
Co Chair OASIS SCA Assembly TC.
IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great Britain.
Phone & FAX: +44-1962-818014 Mobile: +44-7802-467431
Email:
ashok malhotra <>
21/07/2008 23:03
Please respond to
To
Mike Edwards/UK/IBM@IBMGB
cc
OASIS Policy <>
Subject
Re: [sca-policy] [ISSUE 15] External
attachment - Alternative Proposal
Hi Mike:
In addition to the lifecycle issues that Rich has raised, I would like
to bring up a few high-level questions
re. you proposal for issue 15.
1. We can extend the policySet syntax to accommodate policy attachment
but we also need a file format.
This is for the file that contains the policySets and can be included in
a contribution.
<mje>
I note that the current
spec doesn't have this today - and I agree that it needs to be added.
We can either include PolicySets
within <definitions/> (no new files needed...) or we can create
a new PolicyDefinitions
file. I have no strong feelings on this.
</mje>
2. I raised the usecase for corporate policies which, under the existing
specification could be merely included at some
high level and then pulled in as needed. Under the new proposal this
needs a bit more work. I would like to see
the details for this fleshed out.
<mje>
Not sure I see the problem
here.
Take the corporate policies
and contribute them to the Domain.
Provide some corporate
level attachment definitions for those policy files, like
//service/binding.ws
//reference/binding.ws
//service/binding.ws/IntentRefs("confidentiality")
</mje>
3. Policies on interfaces
a. interface.wsdl is identified by a URI that points to a definitions
file. If we want to attach policies to individual interface
elements such as a operation or a message, we should use the WSDL
fragment identifier syntax as used by binding.ws.
<mje>
I proposed using some functions
for this
- InterfaceRef
- OperationRef
- MessageRef
- sounds like you're not
happy with this approach
How would the fragment
identifier proposal work? Do you have some examples?
In what ways is it better
than my proposal?
</mje>
b. A common usecase may be "apply this policy to this binding and
this
interface". This would require an XPath function
that identifies a wsdl interface which has as a sibling a particular
binding. This is not a simple XPath function to
write. We need an existence proof.
<mje>
Any peer based stuff is
hard to write in XPath - but I find it hard to believe that
it is not possible to write
one.
You're asking for a specific
example, right?
</mje>
c. In the above usecase, what does the policy get attached to? The
binding or the interface element or some combination?
If it is attached to the interface it applies to all sibling bindings,
which may not be what we want.
<mje>
Logically it gets attached
to the interface, operation or message as used by the
specific binding identified
(if any).
The rule must be that the
binding gets picked first - and that if any binding is OK, then
//service is used rather
than //service/binding.ws (etc)
</mje>
You will recall that this is the reason Mike Rowley wanted to munge the
domain interface in unnatural ways. But we still
need to solve the problem.
<mje>
Agreed - and I think that
the proposal does solve that
</mje>
All the best, Ashok
Mike Edwards wrote:
>
> Folks,
>
> Here is the alternative proposal for Issue 15 that we have been
> working on:
>
>
>
> Yours, Mike.
>
> Strategist - Emerging Technologies, SCA & SDO.
> Co Chair OASIS SCA Assembly TC.
> IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great Britain.
> Phone & FAX: +44-1962-818014 Mobile: +44-7802-467431
> Email:
>
>
> ------------------------------------------------------------------------
>
> /
> /
>
> /Unless stated otherwise above:
> IBM United Kingdom Limited - Registered in England and Wales with
> number 741598.
> Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire
PO6
> 3AU/
>
>
>
>
>
>
> ------------------------------------------------------------------------
>
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail. You may a link to this group and all your
TCs in OASIS
> at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail. Follow this link to all your TCs in OASIS at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU
