[JDEV] Re: Extra namespaces for legacy protocols?

GuruJ GuruJ at mbox.com.au
Mon Jun 16 05:27:36 CDT 2003


Rachel Blackman wrote:
> The Jabber dev community's time is better spent improving Jabber.  Don't
> look at how to emulate and map the existing IM systems into Jabber; if you
> do that, you'll constrain yourself by what they do.

Yes, yes, yes!  The only way that we'll get people to use Jabber is if 
it offers something substantially over and above that offered by the 
other IM clients.

We may not yet have as broad a community base as something like 
OpenOffice.org or Mozilla, but we have similar advantages in that we all 
use open, extensible formats and protocols that are conducive to 
community extension and development.

Eventually Jabber needs to evolve beyond the X+1th IM Jabber client 
which clones AIM, MSN, or whatever, and provide *substantial* additional 
functionality not provided elsewhere.  It is capable of being *so* much 
more than just an IM system, and I'd prefer to see people concentrate on 
big innovations rather than re-implementing the same technology in place 
already elsewhere.

> What I /want/ to do, eventually, as I said is to sit down and start writing
> JEPs for all the features the other IM systems have.  Not to directly map,
> but to look at what users want from those features.  To implement clean,
> understandable JEPs which provide similar /or preferably better/
> implementations of the features.  Ideally even to write reusable client
> libraries for some of the more complex stuff (Jabber voice chat, whatever),
> to release out there freely.  I'd like to sit down with other authors, like
> the GAIM team and the Psi team and so on, to come up with these JEPs and
> libraries.

Also a really good idea!  I can't help wondering whether it would be 
really worthwhile to implement a generic JEP that specifies the standard 
for writing *client* components -- something along the lines of Mozilla 
XPCOM.

While I acknowledge the risks in pointing to a project that took 5 years 
to really hit its straps, I think that Mozilla's choice to create the 
XPCOM technology has provided a fertile ground for developers that 
promotes the idea of code sharing and re-use.

I know that there are many libraries out there that implement basic 
Jabber functions:  Jabber-Net, pyxmpp, Net::Jabber, etc.  However, why 
re-invent the wheel for higher-level functions over and over on each 
client, or through each library?  A 'lingua franca' that can be spoken 
by *all* languages to access functions implemented by developers would 
be invaluable.

Regards,

GuruJ.





More information about the JDev mailing list