Re: [libreplanet-discuss] Google to drop XMPP entirely

[David Loyall]

On Tue, May 28, 2013 at 9:15 PM, Sergio Durigan Junior
<sergiodj@riseup.net> wrote:
>
> On Tuesday, May 28 2013, David Loyall wrote:
>
> > Google knows what percentage of Talk traffic came from external XMPP and
> > what percentage came from gmail.  Surely they also know that their
> > recent
> > action will cause at least a small surge in global XMPP usage and a
> > corresponding (small) drop in G Talk usage.
>
> Or maybe they are counting that, inside the small group of people who
> use XMPP/Jabber, there will be a large percentage who will "migrate" to
> the Hangouts thingy.  But my point is: they don't care about that.

I read this differently: years ago, they devoted significant resources
to encouraging folks like me (who had never heard of XMPP before
Google Talk) to set up their own XMPP services.  Maybe they envisioned
a healthy federated ecosystem of XMPP domains.  Then, it never
happened.  Maybe users too many users like me tried and failed to set
up XMPP on our own domains, or succeeded but ended up using the
various Google Talk clients anyway because of superior features.

So today, big G might be thinking: you can lead a horse to water, but
you can't make her drink.  Why should they bother adhering to an open
protocol when we weren't using it?

So you're right that "they don't care", but I think it's "they don't
care about supporting XMPP" because even if they were to continue
support, it would only be utilized by hobbyists.

> > I compare this Google XMPP situation with the situation that caused the
> > founder of identi.ca to say at LibrePlanet 2013 that he wishes more
> > people
> > were using federated StatusNet services, rather than all piling on the
> > reference implementation.
>
> Unfortunately I couldn't attend LibrePlanet 2013.
>
> I know federated services are good not only for the web but also for the
> people (taking privacy into account, among other things), but IMO the
> reason people use the "reference implementation" is because (a) they
> were taught that their computers are clients and there must be servers
> (which are *not* their computers!), and (b) setting up a federated
> service frequently is not as trivial as one might wish.  (yes, pump.io,
> I'm looking at you too).
>
> IOW, what Evan says is fine, but it is unfortunately very far from the
> reality.  Take StatusNet and his recent new project called pump.io, for
> example.  What he actually did was to set up several "Try it!" servers
> which will ultimately be used by whoever wants to use the service.  He
> is providing the federation along with the service, maybe just because
> he wants to be able to say later that pump.io is different from
> StatusNet...
>
> Sorry about drifting away from the subject, but I couldn't help myself.
> This is an important topic and I think it should be discussed more.

Sorry, I don't understand the history well enough to parse this.

> > Google could regain some trust from the community by releasing their
> > XMPP
> > stuff as free software.  (If they're not using it, why not, right?)  It
> > wouldn't surprise me if that goal is already on their to-do list, as a
> > low-priority.  (If they are anything like other organizations, they
> > won't
> > ever do any low priority task unless it gets promoted to medium
> > priority,
> > for example, due to user pressure.)
>
> Hm, I didn't know there are Google's XMPP stuff not released.  Do you
> happen to have an example?

I presume that they have a robust XMPP daemon of some sort, which is
about to fall into disuse.

Hm, maybe they could also regain trust by liberating Hangouts
source--server side, preferably.

Note: I am assuming that they want us to like them.  (A common human
trait.)  It might be a pretty safe assumption:
http://www.fsf.org/patrons

I am also assuming that they will either read this, or read some parts
of it that may or may not get incorporated into the open letter.

> Cheers,
>
> --
> Sergio
>
>

Cheers,
--Dave