Re: [sca-assembly] ISSUE-137: Clarify the use of a binding name ina reference target - proposal (c)

[Simon Nash <oasis@cjnash.com>]

Dave,
If we are going to make these bullets parallel, I think we need to
go a bit further than you are suggesting.  I think the second bullet
would become something like the following:

  <service-name> is the name of the target service within the component.

  If <service-name> is present, the component service with @name
  corresponding to <service-name> MUST be used for the wire. [ASM600xx]

  If there is no component service with @name corresponding to
  <service-name>, the SCA runtime MUST raise an error. [ASM600xx]

  If <service-name> is not present, the target component MUST have
  only one service with an interface that is a compatible superset of
  the wire source’s interface and satisfies the policy requirements
  of the wire source, and the SCA runtime MUST use this service
  for the wire. [ASM600xx]

   Simon

David Booz wrote:
> Simon,
> 
> Thanks for clarifying your point. I was certainly aware of the impact of 
> this change. In the past we didn't have interface compatibility rules on 
> which to base this sort of capability, nor did we have the effects of 
> policy well understood.
> 
> Taking into account the subtle nature of this, I'm now wondering if I 
> should change the second bullet (which describes <service-name>) to an 
> RFC2119 statement, similar to the last paragraph of the 3rd bullet.
> 
> Dave Booz
> STSM, BPM and SCA Architecture
> Co-Chair OASIS SCA-Policy TC and SCA-J TC
> "Distributed objects first, then world hunger"
> Poughkeepsie, NY (845)-435-6093 or 8-295-6093
> e-mail:booz@us.ibm.com
> 
> Inactive hide details for Simon Nash ---07/01/2009 04:53:04 PM---Dave, 
> I'm sorry that we ran out of time on Tuesday's call befoSimon Nash 
> ---07/01/2009 04:53:04 PM---Dave, I'm sorry that we ran out of time on 
> Tuesday's call before we could
> 
> 
> From:	
> Simon Nash <oasis@cjnash.com>
> 
> To:	
> David Booz/Poughkeepsie/IBM@IBMUS
> 
> Cc:	
> sca-assembly@lists.oasis-open.org
> 
> Date:	
> 07/01/2009 04:53 PM
> 
> Subject:	
> Re: [sca-assembly] ISSUE-137: Clarify the use of a binding name in a 
> reference target - proposal (c)
> 
> ------------------------------------------------------------------------
> 
> 
> 
> Dave,
> I'm sorry that we ran out of time on Tuesday's call before we could
> finish the discussion of policy matching for target URIs.
> 
> In an attempt to save time on next week's call, I'll summarize my
> point here.  I hope this clarifies things rather than making the
> discussion even more confused!
> 
>  From the spec words before proposal 137c, it was only valid to omit
> the <service-name> if there was exactly one compatible service
> on the component.  So, if a component MyComp has services "Foo"
> and "Bar", and the interfaces of "Foo" and "Bar" are both compatible
> with the reference interface, it would be necessary to specify
> either "MyComp/Foo" or "MyComp/Bar" explicitly in the reference URI
> instead of just "MyComp".  This is irrespective of whether polices
> match or don't match, because policy matching is not part of the
> definition of interface compatibilty.
> 
> The wiring rules require interface compatibility, and they also
> require matching policies on reference and service.  So if "Foo" has
> matching policies and "Bar" does not, the target "MyComp/Foo" would
> be OK and the target "MyComp/Bar" would be an error.
> 
> The 137c proposal would make one change to this: it would make
> "MyComp" a valid shorthand for "MyComp/Foo" based on the additional
> policy matching rule.  It would leave the other cases as before, with
> "MyComp/Foo" being OK and "MyComp/Bar" being an error.
> 
> If you (or others) interpreted the previous words as allowing "MyComp"
> in this case, then the new words are equivalent and clearer.
> 
> If anyone interpreted the previous words as not allowing "MyComp" in
> ths case (as I did), then the new words would change this.  However,
> I don't object to making this change.  One point in its favour is that
> the rule stated in the new words is consistent with how autowiring works.
> 
>   Simon
> 
> Mike Edwards wrote:
>  >
>  > Folks,
>  >
>  > This looks good to me.
>  >
>  > +1
>  >
>  > 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:  mike_edwards@uk.ibm.com
>  >
>  >
>  > From: David Booz <booz@us.ibm.com>
>  > To: sca-assembly@lists.oasis-open.org
>  > Date: 29/06/2009 21:05
>  > Subject: [sca-assembly] ISSUE-137: Clarify the use of a binding name in
>  > a reference target - proposal (c)
>  >
>  >
>  > ------------------------------------------------------------------------
>  >
>  >
>  >
>  > Attached is the updated proposal for ASSEMBLY-137. I believe all
>  > comments from the last telecon are addressed, and in addition I found
>  > several additional places (as I suspected) which also needed a minor 
> update.
>  > /
>  > (See attached file: sca-assembly-1.1-spec-cd03+issue137c.doc)/
>  >
>  > [1] _http://www.osoa.org/jira/browse/ASSEMBLY-137_
>  >
>  > Dave Booz
>  > STSM, BPM and SCA Architecture
>  > Co-Chair OASIS SCA-Policy TC and SCA-J TC
>  > "Distributed objects first, then world hunger"
>  > Poughkeepsie, NY (845)-435-6093 or 8-295-6093
>  > e-mail:booz@us.ibm.com[attachment
>  > "sca-assembly-1.1-spec-cd03+issue137c.doc" deleted by Mike
>  > Edwards/UK/IBM]
>  > ---------------------------------------------------------------------
>  > 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/
>  >
>  >
>  >
>  >
>  >
>  >
> 
> 
> 
> ---------------------------------------------------------------------
> 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 
> 
> 
>