Direct IM requests and the Request API

Evan Schoenberg evands at
Wed Apr 25 12:25:37 EDT 2007

On Apr 25, 2007, at 11:26 AM, Ethan Blanton wrote:

> Looking at these parameters, they seem redundant to me; if a
> conversation exists, the account and who parameters can be derived
> from it, and the account and who parameters can be used to find it.
> If a conversation with that buddy does not exist, the account and who
> parameters can be used to verify this.  It seems that the conversation
> parameter could (and should?) be removed.

I initially did just account/who, but I was trying to work forward  
toward a future in which we aren't so tied to the account + who ==  
conversation model, as that is very one-to-one IM centric.  Including  
the conversation allows associating requests with a particular group  
chat, conference, or IM, if desired.  So in the situation in which  
you have:
1. Bob talking to Jon in an IM
2. Bob talking to Jon and Anne in a group chat

account + who has to assume that we're talking about (1), but what if  
the request is in some way tied to Jon within the group chat (2).

I don't know of a situation offhand in which that's possible with our  
current group chat implementation -- things in the group chat which  
are targeted at a particular user have to be taken out of the group  
chat, into a private message of some sort -- but it seemed like  
allowing for that future expansion could be useful.

Is that too forward-looking for a current API change?  We certainly  
could go back and add that parameter later if it becomes necessary,  
but it'd be time consuming to do later whereas doing it all at once  
now is the same trouble as adding the two parameters.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the Devel mailing list