Concerning internalObjectID...

[Evan Schoenberg, M.D.]

On Aug 16, 2011, at 1:26 PM, Colin Barrett wrote:

> So we have
> 
> - (username, hostname) which won't change without sign off / sign on, but doesn't uniquely identify a user, and
> - nick, which can change without sign off / sign on, but does uniquely identify a user
> 
> Essentially we should create our own internal identifier (just a number) as a replacement for nicks, which will be assigned once when we first see a particular contact and never change. When we see nick changes, we update the mapping of nick -> id num. (perhaps stored with a map of user at host to an array of (nick, id) pairs.)
> 
> Sorry this is just a sketch, but hopefully it can spark some ideas for someone else.

How does this interact with logging (which seems somewhat broken for the case of changing nicks, anyways, so maybe we don't care)?

-Evan

> 
> -Colin
> 
> On Aug 16, 2011, at 1:16 AM, Thijs Alkemade wrote:
> 
>> The servers I'm familiar with (Freenode, Quakenet) all make you appear to part or quit when you apply a cloak.
>> 
>> However, without nick it is not unique, which would give the same problems as non-anonymous XMPP MUCs, where we can not distinguish different instances of the same JID. There too we should find a better way to ID users.
>> 
>> Thijs
>> 
>> On Aug 16, 2011 2:50 AM, "Colin Barrett" <colin at springsandstruts.com> wrote:
>> > On Aug 15, 2011, at 5:29 PM, Peter Hosey wrote:
>> > 
>> >> username at hostname is not changeable except by disconnection and reconnection, but multiple connections to the server can have the same u at h (with different nicknames).
>> > 
>> > Except on freenode when you authenticate with their services. You can have your hostname changed at that point to a vanity/privacy mask.
>> > 
>> > -Colin
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://adium.im/pipermail/devel_adium.im/attachments/20110816/6dd14865/attachment-0002.html>