MeatballWiki | RecentChanges | Random Page | Indices | Categories

Page Accueil
http://www.jabber.org (autrefois jabbercentral.org)

Jabber est à première vue un système OpenSource de MessagerieInstantanée?. Il a démarré initialement comme un moyen de connecter ensemble tous les grands systèmes de MessagerieInstantanée?. Mais il est devenu beaucoup plus cool que ça. C'est une couche de transport push. Il est fondé sur XML.

http://www.jabber.com est la face corporate du projet jabber (Jabber, Inc.)


Commentaires / Discussion

It's really good. -- Sunir@jabber.org

Mais il a des problèmes politiques. -- SunirShah

Puis-je demander la nature des problèmes politiques avec Jabber, ou leur expliquer trop de politique ? Et pour commenter, je comprends que les ProtocoleOuvert?s sont plus importants que l'OpenSource, raiso pour laquelle les RFCs sont importantes. -DaveJacoby

Oui, ce n'est pas machiavélique, mais c'est politique. Jabber.com contrôle le projet. Ce n'est pas un projet ouvert. L'open source produit un programme ouvert, mais ne produit pas un projet ouvert. Seuls des choses comme le ProcessusEquitable le font. -- SunirShah

Dave, ceci peut expliquer quelques points. Gardez à l'esprit que c'est vraiment un manque de temps qui engage Jeremie Miller, le leader jabber.org. Ayant discuté avec lui, je crois qu'il sait ce qu'il fait. D'un autre côté, je ne crois pas que Jabber.com a de bonnes intentions. (P.S. je me sens à l'aise pour poster cet email qui a déjà été distribué au conseil open source Jabber.com.)

From:	Sunir Shah [sunir@sunir.org]
Sent:	March 1, 2001 3:54 AM
To:	'esr@thyrsus.com'
Subject:	RE: Jabber open board and its viability
> Aren't they releasing source?  If they are, then in what sense
> are the cards close to the chest?

Released source code doesn't make a *project* open, merely the
program. To make the project open, input from the userbase must
be accepted. Even better, the distinction between the project and
the userbase should become fuzzy. That is, I as a user should be
able to do development when the program does not meet my
requirements. Moreover, since this application has no value by
itself, but as a network, my changes must be accepted by a
significant portion of the users to provide value. Without
entering the core stream, this is next to impossible.

The Jabber project is currently developed soley by a handful of
core developers. Project planning is downstream only;
announcements are made on http://dev.jabber.org as to what they
are doing. But there is [no] feedback channel to get things integrated
into the release schedule short of petitioning them on the jdev
groupchat on jabber.org.

Furthermore, as Jabber.com owns all the code, there is a
significant wall to upstreaming patches to the open group. They
have a stable version of the server in house, yet the open
version is not (*). There is little reciprocity. Worse, they have
tied up all the open source developers with their internal

[(*) A rather poor decision as many people have concluded that
Jabber isn't working, as the open server hasn't been stable for
awhile. Specifically, the transports to other instant messaging
networks like AOL Instant Messenger, etc. have not worked for a
long time, and they are Jabber's raison d'etre. For another
famous example of this, consider Mozilla. Many people still won't
touch it because it didn't work early on.

The significant reason for the unstability has been lack of
developer resources. All the users get are messages that "we're
working on it," but the jdev mailing list has been empty of
information on what's broken. If there were more appeals for help
outside the inner circle, I'd bet things would be better.]

Consequently, the project is not open. And I believe this is
because of pressure from Jabber.com.

I believe this is what you originally referred to as Cathedral
development a long time ago, if I understand correctly. When you
were comparing two differen open source development styles, not
open vs. closed source.

Have you ever heard of a philosophy called "Fair Process?" I have
put some comments at 


An open ended organization like an open source project must
maintain open communications, plus the leadership must be 
attentive to the constituency, or the constituency will perceive
(correctly) a lack of procedural justice. Thus they will bail.


A personal anecdote on my experience with Jabber (extremely limited though it may be). KuroShin was one of the first places any news about Jabber appeared, in any medium. A couple of the kids who were working on it wrote up a great little article [1] back in May 2000, which sticks in my memory because they went through about three or four revisions of it, and I was impressed by the lengths they were willing to go to to get it right. And read the article: they really care about it's being open source.

Skip forward to LinuxWorld Expo, in San Jose, circa August 2000 (?). I was wandering the show floor, and ran into a couple guys wearing Jabber shirts. They were mid-fortys-looking, and looked very much like Management. However, I had no idea who "Julian" was, other than having conversed a couple times on IRC and via email. I said hello, and told them who I was.

"Rusty. From Kuro5hin."

[Blank looks from the Managers]

"We ran an article on you way back in May..."

[More blank looks]

"Is Julian here?"

Finally some light dawns. Yeah, they tell me, he's around somewhere. They wander off, obviously not giving a shit who I am.

A while later, I run into a couple kids who look about 13, also wearing Jabber shirts. I was hanging out at Linux Journal with Doc, and they came right up to me.

"Hey, are you Rusty?"


"I'm Julian, and this is [forget the name, d'oh]. We heard you were over here."

Pleasantries were exchanged, they were really happy to meet me, etc. I eventually asked them who the Management dweebs were. "They're the .com guys" was pretty much the answer. Jabber was bought by some company, and was clearly operating with two very different camps. I heard, in the back of my mind, a deep minor-key note that boded ill for the future.

It looks like what you're seeing now is the upshot of the basic disconnect between the very-committed-to-free-software Jabber hackers, and the very-committed-to-making-a-buck .com parasites. I'm not at all surprised to hear this, and I hope they can weather incompetent management and come out a stronger project. Watching the same basic mismanagement occur at my last company [2], I'm kinda sensitive to the symptoms now.

Wow, that was a lot longer than I thought it was going to be. :-)


I was just talking with JeremieMiller?, leader of Jabber.org. Jabber.com and Jabber.org just went through a significant reshuffle of its priorities. Jabber.org will be improving its developer focus now, trying to open up (yes, because of my e-mail). What I would think might be interesting is plugging something like a real time Jabber-based "WhoIsOnline" system into a site, or a network of sites. Or even updating a network of sites in real-time with each other's status using their dedicated socket protocol. Jabber is a lot more than InstantMessaging, like it's said above. -- SunirShah

Since Jabber is an open protocol, anyone can steal the entire project by implementing a better server or client that becomes the new standard. Sort of like IE took over the HTML standard. For this reason, I think Jabber is still good. --anon.

That's not true at all. There is more to supplanting the standard than making a technically better server. In fact, you don't even need a technically better server, just a more popular one. In fact, look to IE for an example of that. Certainly IE was not technically better than NetScape when it began taking market share. -- SunirShah

A bit off topic, "[forget the name, d'oh]" was me. -- [EliotLandrum]

LangueFrançaise PageTranslation JabberProject


MeatballWiki | RecentChanges | Random Page | Indices | Categories
Edit text of this page | View other revisions | Search MetaWiki