[standards-jig] thoughts on a new idle protocol

Dudley Carr dudley at cs.stanford.edu
Sun Jan 18 20:48:45 UTC 2004


Bob Wyman wrote:

> Dudley Carr wrote:
> 
>>Most people think of idle time as the time since the person 
>>last moved the mouse or typed in their IM application. IMO, 
>>a more sensible approach is to measure idle time starting 
>>when the last away presence stanza was sent (automatically or 
>>manually).
> 
> 	Personally, I think that idle should be an indication of what
> the person has been up to -- not what the machine serving that person
> may or may not have done recently. Thus, I would not support your
> interpretation of idle... Nonetheless, the Pubsub protocol can
> accommodate both... 

My interpretation is consistent with your definition of idle reflecting what the 
  person has been doing, but with a single additional constraint: my computer 
shouldn't be telling others that I'm here (not idle) when I've already said that 
I'm away.

AIM does this where you've set an away message but if you're still browsing the 
web, the people on your buddy list still now that you're there. I guess it's 
arguable that some people may want this behavior.

> 	If we say that there are different "styles" of idleness, then
> you can interpret this as a question of PubSub subscription
> configuration. For instance, define two styles:
> 	1. PersonPresent
> 	2. TimeSinceLastAway
> 	Then, provide a subscription configuration option that allows
> the subscriber to select which of the two styles they are interested
> in. 
> 

The PubSub solution is a fine one, but it seems to me that you can already 
figure out all the idle information from received presence stanzas.






More information about the Standards mailing list