[Members] http://xmpp.org/2011/06/skype-adds-xmpp-support/ <EOM>
dave at cridland.net
Wed Jun 29 10:47:53 UTC 2011
On Wed Jun 29 07:05:46 2011, Waqas Hussain wrote:
> The other post on GigaOM is better, and says the opposite for
> Microsoft: "Microsoft’s own Windows Live Messenger currently
> support XMPP"
Right. Microsoft's enterprise IM server, however, does support XMPP
for S2S, and quite well these days from what I hear.
> Yahoo and others have experimented with XMPP, and any federation
> between these networks will likely use XMPP. That said, I think
> open XMPP use is unlikely for any of these services, at least in the
> next couple of years. There just isn't much incentive for it.
As you hint, it's the S2S that's most likely to open up, to the
extent that it's reasonable to refer to it in terms of a foregone
conclusion. I think there is incentive to do this - if not now, then
in a matter of months - because of the increasing number of
enterprise IM systems which already offer XMPP/S2S.
I'll even make a prediction - just as Google Talk turned on open
federation after a few months, so will AOL, and Yahoo will follow
with an XMPP/S2S offering of their own. Skype, too, will open up
basic XMPP/S2S interop, although we'll be moaning that they don't do
Jingle for a very long while.
MSN/Live will hold out longest - they have federation into their own
Enterprise solution, I think, as well as federation with Yahoo if
memory serves, and as such they simply have far less to gain.
Now, as an organisation, we can approach this in two ways:
1) Make the assumption that users using a closed C2S, but an open
S2S, will outnumber "pure" XMPP users, and handle this accordingly in
our protocol designs. In other words, we make much more effort to
devise protocols which will gateway neatly onto existing IM systems.
I'd note that much of the core protocol is designed this way as a
legacy of the beginnings of Jabber, so this is not too hard to do.
2) Whilst acknowledging that XMPP/S2S is used as a gateway protocol,
we move to concentrating on demonstrable benefit to pure XMPP users.
This might include Jingle and other end-to-end technologies, but it
may also involve non-IM use-cases over the IM network.
We don't have to just pick one, incidentally.
An example of "Option 2" being put into action would be the Jingle
Interop event mentioned a few weeks back. An interop event serves two
purposes - it allows developers to test against a wide number of
clients with their fellow developers ready to help debug what goes
wrong, and that means a much shorter find-fix-test cycle. But it also
makes for a very clear demonstration to outside people that the
technology works and has interested and active developers, and this
is also vitally important.
In the case of a Jingle Interop, what we'd be doing is demonstrating
that actually, we have the whole Skype thing covered.
Dave Cridland - mailto:dave at cridland.net - xmpp:dwd at dave.cridland.net
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade
More information about the Members