Peter Saint-Andre wrote:
> Sebastiaan Deckers wrote:
>> On Tue, Oct 14, 2008 at 2:12 AM, Peter Saint-Andre <[EMAIL PROTECTED]
>> <mailto:[EMAIL PROTECTED]>> wrote:
>>
>> 1. Who has implemented XEP-0012? Please note that the protocol must
>> implemented in at least two separate codebases.
>>
>>
>> I implemented this for Pandion a long while back. Client requests the
>> iq:last time to show the user when a currently offline contact was last
>> available. This provides useful information in many cases.
>>
>>
>> 2. Have developers experienced any problems with the protocol as defined
>> in XEP-0012? If so, please describe the problems and, if possible,
>> suggested solutions.
>>
>>
>> This specification is very trivial to implement. However the design is
>> inherently pull based, as others have mentioned. In the implementation
>> that I created that's an efficiency benefit, but making it push based
>> would enable many other kinds of uses for this information. PEP anyone?
>
> In fact I think it would be find to push this information in presence.
> Or perhaps a server simply needs to notate last presence with a delayed
> delivery timestamp -- then you would not need iq:last at all, I think.
No, that's wrong. You'd still need the iq:last in presence, which would
be further notated as delayed. This is like the past perfect tense in
grammar: at that time I *had been* idle for 2 hours...
<presence from='[EMAIL PROTECTED]/orchard'>
<show>away</show>
<query xmlns='jabber:iq:last' seconds='7200'/>
<delay xmlns='urn:xmpp:delay'
from='montague.net'
stamp='2002-09-10T23:08:25Z'/>
</presence>
This means that at 23:08:25 UTC on 2002-09-10, Romeo had been idle for 2
hours. The iq:last notation is provided by the client, and the delayed
delivery notation is provided by the server.
Peter
--
Peter Saint-Andre
https://stpeter.im/
