RE: [ws-caf] proposal for issue 243

["Martin Chapman" <martin.chapman@oracle.com>]

Not sure if people saw the notice of the ballot for this issue, but it
closes this Saturday
http://www.oasis-open.org/apps/org/workgroup/ws-caf/ballot.php?id=726

>-----Original Message-----
>From: Mark Little [mailto:mark.little@arjuna.com] 
>Sent: 26 March 2005 08:58
>To: ws-caf
>Subject: Re: [ws-caf] proposal for issue 243
>
>
>I've set up the vote. Separately, I propose adding the 
>following text - 
>I didn't want to put it into the vote in case people agreed 
>with the UTC 
>format, but had issue with the text:
>
> In addition, add the following text:
>
>"Implementations MAY use fine-grained clock synchronization protocols. 
>However, this specification does not require such protocols and 
>therefore timeout values that are specified with seconds MUST NOT be 
>relied upon unless additional information is available which 
>is outside 
>the scope of this specification."
>
>Mark.
>
>
>
>Mark Little wrote:
>
>> Does anyone have any comments on this? I'd like to start a vote on
>> this in the next few days, so please let me know by Friday.
>>
>> Mark.
>>
>>
>> Mark Little wrote:
>>
>>> Actually as Kevin just pointed out to me, this won't work unless the
>>> receiver knows when the duration started. So, I suggest we go for 
>>> dateTime and require it to be UTC. Obviously there'll need 
>to be some 
>>> text about the fact we don't require clock synchronization 
>protocols 
>>> to be run, so this should not be used for fine-grained time 
>decisions.
>>>
>>> Mark.
>>>
>>>
>>> Mark Little wrote:
>>>
>>>> http://services.arjuna.com/wscaf-issues/show_bug.cgi?id=243
>>>>
>>>> With reference to http://www.w3.org/TR/xmlschema-2/#isoformats
>>>>
>>>> I don't really have a preference either way (dateTime or duration),
>>>> but since the context timeout is meant to be the duration for the 
>>>> context, it seems to make more sense to set this to be the 
>duration 
>>>> (http://www.w3.org/TR/xmlschema-2/#duration).
>>>>
>>>> Mark.
>>>>
>