[jdev] GSoC proposal - open web XMPP client

Sander Devrieze s.devrieze at pandora.be
Sun Mar 30 13:43:27 CDT 2008

2008/3/30, Tomasz Sterna <tomek at xiaoka.com>:
> Dnia 2008-03-30, nie o godzinie 17:48 +0200, Sander Devrieze pisze:
> > >  Why another web client? What is wrong with JWChat?
>  > >  And if there is something, maybe it's better to put the efforts in
>  > >  extending/fixing JWChat?
>  >
>  > Because competition can result in a better JWChat client *and* a new
>  > client...and 2 good clients is better than 1 ;-)
> The "competition law" would be true if we were on the commercial free
>  market.
>  But once we are in the Open Source world, I seriously doubt that we have
>  enough manpower to use the competition to our advantage.
>  On the contrary - experience shows that segmentation of efforts in OSS
>  leads to gazilions of half-finished projects.

...at least the Coccinella and the ejabberd projects learned from
projects that died. E.g., ejabberd has some of the features we found
in the WPJabber changelog and Mats fixed compatibility bugs in
Coccinella found when testing things like file transfers with other
clients (some of them now death of semi-death). So yes, I'm in favour
of *more* new XMPP clients, *more* new XMPP servers, and *more* new
XMPP libraries. Advantages of *more*:
* people who start new projects are creative and have interesting
ideas regarding features and interface design-->incumbent projects
will see these sexy features and will copy them (at least we actively
do this at Coccinella, and other projects at least do this passively
by users requesting the same features)
* more people will thoroughly look at the core specifications of XMPP
and thus more bugs and implementation issues in them will be
found-->the XMPP specifications will get better and this will benefit
the whole community
* more compliancy bugs with incumbent projects will be found and these
issues can be fixed. A good example of this is when ejabberd started
to become much more popular where in the past jabberd14 had a
monopolistic market share. When this happened several client and
library projects found out they had implemented some XMPP specs
wrongly.-->with more clients, libraries and servers, the "Web of
Compliancy" will get stronger and it will be hard to be incompliant.
This is a good protection against strategies like this one:

> See sf.net for reference.

How many campanies go bankrupt each year? How many of them go bankrupt
within the first 5 years of their existence. This number is high,
especially in highly competitive industries like the Horeca. Anyway,
in case you are interested you can find some bankruptcy statistics for
Belgium here: http://statbel.fgov.be/indicators/bri_en.asp

Mvg, Sander Devrieze.

More information about the JDev mailing list