On Jul 21, 2011, at 08:39, Peter Saint-Andre wrote:

> On 7/21/11 8:37 AM, Kevin Smith wrote:
>> On Thu, Jul 21, 2011 at 3:28 PM, Dave Cridland<[email protected]>  wrote:
>>> On Wed Jul 20 23:00:33 2011, Matthew A. Miller wrote:
>>>> 
>>>>> If there is only one resource online and the resource updates presence
>>>>> but remains available.
>>>>> 
>>>> 
>>>> This seems reasonable.
>>>> 
>>>>> If the locked resource updates presence but changes only<status/>.
>>>> 
>>>> I don't agree with this one.  There are clients that, most often at the
>>>> explicit direction of their users, only update<status/>.
>>> 
>>> Well, I think that's the point, and it's fine to stay locked. If the change
>>> of any informational aspect of presence has to cause unlocking, then you
>>> should be mandating unlocking on several PEP events, too.
>>> 
>>> But we don't have to agree, here - just give the first case above as a "for
>>> example", and note that it's always safe to unlock.
>> 
>> I'll go further than this, I don't think we can claim it's a best
>> practice to unlock on text-only presence changes.
>> 
>> I'd suggest something along the lines of: SHOULD unlock on any change
>> of availability (e.g. changing from away to available) and MAY unlock
>> on other presence changes (e.g. where the text changes).
> 
> Sounds right to me.
> 

No, sorry, it's not right at all.

Like I said, there are a significant enough number of clients/users that will 
use status and only status to indicate change in availability.  Whether that's 
right or not can be debated, but this line **WILL** break behavior for a number 
of people, and there's no sure fire way to detect such behaviors.


- m&m


Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to