[Members] Google and XMPP federation

Dave Cridland dave at cridland.net
Mon May 27 13:48:05 UTC 2013


On 27 May 2013 12:04, "Simon Tennant" <simon at buddycloud.com> wrote:
>
> Which exact statement by Google is technically wrong?
>

That XMPP fundamentally cannot be used to solve their use-case?

One of the problems here is that their statements are either vague, or else
made in proxy by journalists - "Google told us that" statements, I mean.

This doesn't help anyone, of course, but it hurts us precisely because
they're very broad statements coming out.

> So by all accounts we cannot do group communication with file-sharing and
video and reliable message delivery across multiple platforms and any
solution would have to do UX workarounds to account for why some users
don't get a file or perhaps might not receive a message. But Google should
have used XMPP.
>

The "reliable message delivery" is yours. Even Google haven't said that. We
do know that multiparty filesharing and A/V are shortfalls in the spec,
we've known this for years and been actively working on it off and on.
Google have been most active in the community when talking about Jingle.

However, these *can* be solved - it's hardly even a difficult problem.

> Also, reading these comments there's an expectation of "you use XMPP and
you should publish any subsequent extensions". (did Stallman just join the
XSF?). This is new to me. It's also a very scary prospect and will only
serve to scare off new companies investigating XMPP as part of a comms
strategy.
>

No, that's misreading.

Google have claimed to have pioneered XMPP, and they've also claimed it
will not and cannot be used to solve their existing use-cases.

> We're sorely misguided if we think that chastising a relatively good
champion of XMPP for many years (compared with complete non-federators like
WhatsApp and Facebook) will help the cause of open communications.
>

Again, I think you're misreading. Nobody's claiming that Google haven't
done a lot of good by using XMPP in the past - I am, however, arguing that
they're rapidly undoing a lot of this good with the noises coming from that
direction now.

> If we are serious about encouraging open, federated communication and
building the standards that others can adopt then now is the time to lay
down our pitchforks and reflect on:
>
> - why so many web developers have such a negative reaction to XMPP,

Because it's an XML based protocol running over a stateful TCP port - we're
not going to change that. But nobody - I think - is arguing that all
messaging must be XMPP. The key thing we'd like is for federation and
interop with third-party clients (and servers). What happens in a web
browser can stay in a web browser.

> - why XMPP is perceived as difficult to build on (keyword: perceived),

Bare in mind that Google is really helping this particular perception. This
*is* something we can tackle.

> - what we are doing to grow beyond pure user to user chat and better
address the mobile world.

Simon, I know that when you tried basing buddycloud on MUC 5 years ago or
so on Symbian, it didn't work out. I think many things have changed in both
mobile and XMPP capabilities since then, and I think it'd be much easier to
maintain backwards compatibility now and achieve your goals.

That all said, there are things we could do to make XMPP on mobile more
in-tune with current thinking - for example, I was wondering whether
writing a generic mobile push request service might be useful. (Quite
honestly, 198 + "Come reconnnect" is all we need).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.jabber.org/pipermail/members/attachments/20130527/0394a8ef/attachment.html>


More information about the Members mailing list