[JDEV] Best way to drive Jabber adoption?

[Justin Karneges]

On Saturday 14 June 2003 02:18 am, Timothy Carpenter wrote:
> do that, I suggest we have each desktop run a small private SERVER that has
> all the transports.

The problem is that the current Jabber transport software is notable for being 
much less featureful than the multi-IM clients out there.  If you could make 
all of the transports super-featureful with file transfer and everything 
(which I guess is fine for local use) then you basically would end up with a 
multi-IM library that uses XMPP for IPC.  My gut says this is overkill, but 
it would certainly work I guess.

> The server can use s2s to connect to the 'proper' Jabber world (or even a

Linking this embedded jabber server to the real world would not be possible, 
as far as I can tell.  The client could simply maintain two Jabber 
connections (one to a real Jabber server, and optionally one to the local 
server).

> Client designers can then focus on look-and-feel, skins, usability,
> thinness and robustness and forget about having to handle multi-platform
> issues which, frankly, each client is re-inventing the wheel in that
> regard.

They already do focus on these things.  I don't think many clients out there 
are trying to do multi-IM, of which I can only name a handful.  I think 
multi-IM clients help Jabber's world domination agenda, by allowing people to 
easily migrate.  However, if _every_ client was multi-IM, I think we'd be 
putting too much emphasis on legacy systems.

Would it be good if our super friendly client for end-users worked natively 
with AIM/ICQ/MSN/Yahoo?  I'm not sure, but it is something to think about.

Do we need a super portable embedded jabber server transport bundle thing?  
Probably not, regardless of the answer to the previous question.

-Justin