[JDEV] Best way to drive Jabber adoption?
GuruJ
GuruJ at mbox.com.au
Thu Jun 12 03:22:13 CDT 2003
I know heaps of people have already commented on this, but personally I
think Jabber's biggest strength (*at the moment*) is in Jabber networks
that run behind company firewalls.
That's how I first learnt about Jabber, and it's still the killer app
for me as an IT specialist. I think people will still seek out and
adopt Jabber to get this very significant benefit.
While it would be lovely to have a consumer-grade IM product out there,
the fact is that centrally hosting such a service (as people have
pointed out) costs heaps of money for bandwidth and server costs.
Additionally, the administration, maintenance, moderation and content
creation time would be almost prohibitive on a volunteer network.
Finally, as long as interoperability remains a distant dream, there will
always be great difficulty in converting people to Jabber. Why use
Jabber with a MSN gateway (that may or may not work on any given week)
when you can just use MSN?
So, how does a relatively small organisation such as JSF and its hordes
of volunteers compete against the behemoths of AOL, Yahoo! & Microsoft?
I think the Jabber community has three main options:
(1) Set up a centralized jabber.net website. Pick an 'official' client.
I understand that agreement on the 'best' of anything is almost
impossible for an open-source community, but it's *essential* for
end-users to feel comfortable that a 'sanctioned' client will do what
they need, and be well-supported. As noted above, this may be very hard
to fund properly.
(2) Ignore the consumer market and continue to drive adoption in
company-wide IM systems. Once IM interoperability becomes critical mass
and/or more funding becomes available, move to a more consumer-oriented
model.
(3) Come up with a way for consumers to connect to Jabber-based systems
through some kind of distributed mechanism (think Gnutella or
BitTorrent) that only *authenticates* centrally at jabber.net. In this
way, multiple small servers could potentially serve a very large user
community. Of course, the big problem here is that the Jabber protocol
doesn't make provision for this kind of system ... yet.
What does everyone think?
Regards,
GuruJ.
More information about the JDev
mailing list