Re: [sca-assembly] Exit Criteria Proposal

[Jeff Mischkinsky <jeff.mischkinsky@oracle.com>]

On Jun 14, 2011, at 5:36 AM, Martin Chapman wrote:

> Regarding 2) and with a Chair's hat on, I don't think we need to  
> build proof into the exit criteria as I assume TC members might  
> actually want to see some proof. We do however need an deterministic  
> way to determine whether the exit criteria has been met, and that  
> IMHO would be via a motion.

Here are the words from the charter:
The TC shall define concrete exit criteria that include at least two  
independent offerings that implement and are compliant with the all  
normative portions of specifications and demonstrate interoperability  
and portability as appropriate. Note that these are minimums and that  
the TC is free to set more stringent criteria.

I guess i'm at a loss as to why we think the first part (two  
independent offerings that implement and are compliant with the all  
normative portions of specifications) needs to be rewritten. Is it not  
clear?

I actually think the harder part is defining what the TC will do about  
the second part (demonstrate interoperability and portability as  
appropriate). In my book a statement from from somebody saying "sure  
we do it" doesn't constitute "demonstration" of anything.

When I wrote those words it seemed obvious at the time that  
implementers the specs would of course use the TC's tests -- why  
wouldn't you use a test suite for regression purposes that someone  
else wrote?. Of course at the time the assumption was that there would  
be lots of implementations and that the discussion wouldn't be around  
trying to interpret the words to somehow lower the minimum bar.
So i think the TC should discuss how interoperabilty and portability  
should be "demonstrated" if the TC's tests are not used.

cheers,
   jeff
>
> Martin.
>
>
>> -----Original Message-----
>> From: Anish Karmarkar
>> Sent: 14 June 2011 00:32
>> To: sca-assembly@lists.oasis-open.org
>> Subject: Re: [sca-assembly] Exit Criteria Proposal
>>
>> [strange: didn't get Ashok's email, but got Eric's reply and see it  
>> in
>> the archives]
>>
>> I think we ought to have the exit criteria where the two (or more)
>> implementations are conformant therefore each implementation  
>> implements
>> *all* the mandatory normative statements. That is what the spec is
>> asking the implementers to do and the whole idea of the exit  
>> criteria is
>> to ensure that what we have asked the implementers to do is
>> implementable, reasonable, and not an ivory-tower experiment.  
>> Sometimes,
>> two features when implemented together make things very difficult or
>> result in conflicts, or creates issues that don't occur in  
>> isolation --
>> and we should know that
>>
>> There are two more facets of the exit criteria that we are not  
>> included
>> here:
>>
>> 1) What about optional normative statements:
>> Granted the Assembly spec does not have this facet any more since
>> everything is mandatory. Not sure that is true of all other spec.
>> Specifically the policy spec. Whatever we decide here will be  
>> picked up
>> by other TCs.
>>
>> 2) What evidence is the TC going to accept to satisfy itself that two
>> (or more) implementations do in fact implement the spec? The last  
>> time
>> this was discussed, there was a notion that we ought to ask the
>> implementations to run out tests. I believe this is the biggest
>> contentious facet of the exit criteria given the differing views  
>> amongst
>> the TC members.
>>
>> -Anish
>> --
>>
>> On 6/13/2011 3:58 PM, Eric Johnson wrote:
>>> Hi Ashok,
>>>
>>> Which is funny, because I read it as "for each normative statement,
>>> there exist at least two implementations," especially since the work
>>> "every" doesn't appear in Mike's original. That is, Mike could have
>>> written: "there shall be at least 2 independent SCA runtimes that  
>>> are
>>> compliant with /every/ normative portion of the specification..."
>>>
>>> I think Jim's concern is reasonable in theory. In practice, since
>>> Tuscany can claim conformance, as I understand it, then is this  
>>> question
>>> moot?
>>>
>>> -Eric.
>>>
>>> On 6/13/11 3:43 PM, ashok malhotra wrote:
>>>> Jim:
>>>> Mike's proposal says ...
>>>>
>>>>> there shall be at least 2 independent SCA runtimes that are  
>>>>> compliant
>>>>> with each normative portion of the specification as described in
>>>>> Section ...
>>>>
>>>> I read this as saying "two independent runtimes, each of which  
>>>> complies
>>>> with every normative statement in ... "
>>>>
>>>> Ashok
>>
>> ---------------------------------------------------------------------
>> 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
>>
>
> ---------------------------------------------------------------------
> 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
>

--
Jeff Mischkinsky			          		jeff.mischkinsky@oracle.com
Sr. Director, Oracle Fusion Middleware 				+1(650)506-1975
	and Web Services Standards           			500 Oracle Parkway, M/S 2OP9
Oracle								Redwood Shores, CA 94065