Re: [sca-assembly] NEW ISSUE: Inconsistency arising from ASSEMBLY-16changes

Logged as http://www.osoa.org/jira/browse/ASSEMBLY-110

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:  

From:
Simon Nash <>

To:
OASIS Assembly <>

Date:
11/02/2009 23:08

Subject:
[sca-assembly] NEW ISSUE: Inconsistency
arising from ASSEMBLY-16 changes

TARGET: SCA Assembly specification

DESCRIPTION: Inconsistency arising from ASSEMBLY-16 changes

The resolution of ASSEMBLY-16 intriduced the notion of "structural
URI"

as distinct from "binding URI".  The specification had previously
been

unclear on the semantics of URIs for assembly model constructs such

as components, services, references and bindings.

The resolution to ASSEMBLY-16 was in two parts:

  1. A definition of "structural URI" to identify elements
within the

     assembly model, based on their position within the assembly

     structure.

  2. A clarification that the @uri attribute on <binding> specifies

     a "bindingURI" for use by bindings in a binding-specific
manner.

The updates for this resolution were in sections 6.8 and 9.14 (section

numbers are taken from CD02).  Unfortunately some text in section
9

(Binding) was overlooked in the agreed proposal and as a result there

is an inconsistency between section 9 and sections 6.8/9.4/9.4.1.

In section 9, the last sub-bullet under the description of the

binding @uri element says:

  For a binding of a service the URI attribute defines the URI relative

  to the component, which contributes the service to the SCA domain.

  The default value for the URI is the value of the name attribute

  of the binding.

This text implies that only relative URIs can be specified, and that

the resulting URI so defined (...defines *the URI*...) is always

relative to the component.  However, the form of the resulting

resolved URI is not defined.  (The closest we have in the specification

is the form given in section 6.8, but this only uses the binding name

and does not use the binding's @uri attribute.)  This text in section
9

is also not consistent with section 9.4 which says that the bindingURI

specified in the binding's @uri attribute is used in a binding-specific

manner, nor with section 9.4.1 which gives examples of valid binding

URIs that aren't relative to the component, such as "jms:" URIs.

These contradictions result from an oversight in the resolution

to ASSEMBLY-16 which did not adapt the text in section 9 to the

changes made in sections 6.8 and 9.4.

PROPOSAL:

Replace the above quoted words from section 9 by:

  For a binding of a service the URI attribute defines the bindingURI.

  If present, the bindingURI can be used by the binding as described

  in section 9.4.

---------------------------------------------------------------------

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