Ted Leung on the air: Open Source, Java, Python, and ...
I'm going to respond to the thread of comments on the Growl/Zephyr/Jabber posts here, in the hopes that a regular post might help disseminate some of the information in related posts and comments.
Peter Millard, the author of Exodus wrote a post that addresses some of Mark Eichin's criticism of XMPP vs Zephyr. Peter and Mark, if you want to continue clarifying, etc, please feel free to use the comments on this post to do so. Peter was also somewhat put out because my criticism of Jabber clients was not concrete, which is true
Yesterday I asked Peter St. Andre if there was a scorecard comparing the features of the various Jabber clients. Here's why I asked that, and why (in part) my criticism of Jabber clients has been abstract. If you go to Jabber.org and look at "What is Jabber®?", you get a description that talks about open standards, some stuff about servers, security, and extensibility. What is missing, in my opinion, is a clear list of features that the Jabber/XMPP protocol enables, like "headlines". I want to see the client features (like headlines or security) that I don't get (or are hard to get) with other IM systems. Then I want to know which clients actually implement those features. From where I sit, it's very hard to know what is possible with Jabber/XMPP solutions, other than reading a stack of RFCS and JEPs. Unless I can tell people why I (and therefore they) are switching to Jabber, it's hard to make the case. The opening paragraph of the Jabber.org overview is about open standardsness. My own use of IM tells me that while I value open standards, I value features, and my network of people more. I use iChat, not because I love AOL (and I'm happy for whatever Jabber support Apple tosses into Tiger), but for one reason. It's the only client that I can use my iSight with, and the iSight video conferencing is the first web cam style thing that actually works for me. I'm starting to use Skype. I hate the proprietaryness of their protocol. But no one's audio sounds better (not even iChat), and nobody lets me do an n-way voice call. There are like 147 JEPs (okay less than that, but still a lot), most of which were presumably introduced to enable features. But beats the pants off of me if I could tell you what those features are.
I've been writing about Jabber because Lisa Dusseault and Joe Hildebrand tried to educate me one day. They sort of succeeded, because I know that there's a bunch of cool stuff that Jabber can do. But I don't really know what all that cool stuff is, and I have no idea whether a particular client can do the particular cool thing that I need. Jabber is a cool system, that can do cool stuff. I'm convinced of that (at least until Mark and Peter mix it up again). The problem is, I can't actually tell somebody else what any of the cool stuff is. I could read the pile of RFCs and JEPs and eventually figure it out. But it sure would help if someone who already knew could explain it for the rest of us.
Okay, so all that was abstract. Here's a concrete request. Which XMPP Macintosh client should I use if I want to experiment with headlines, so that I could do my system notifications with that instead of Growl?
Here's another one. All you Jabber folks come and leave a comment on your favorite cool - not found in any/most other IM systems - Jabber feature. I'm professing my ignorance -- here's your chance to educate me.
[ After I finished drafting this post, I discovered Dare's posting of the press release on MSN Spaces and the new MSN Messenger, and his personal favorite features. Those of us in the "free world" ought to be able to build stuff like this, and we need a substrate like XMPP to do it. ]
http://www.ietf.org/html.charters/simple-charter.html
It might be fun to work on adding SIMPLE to chandler. There's already SIP support in Twisted (and shtoom). :-)
Take care.
Posted by John P. Speno at Thu Dec 2 05:25:58 2004
I'm not sure what you're interested in doing with headlines on the Mac, but I'm willing to make customizations to Nitro (http://nitro.jabberstudio.org) to help you experiment. I've picked up working on it again, and it may be just the ticket for experimentation.
Drop me a line and I'll see what I can do to help. :)
D.
Posted by Dave Smith at Thu Dec 2 06:28:40 2004
My company, Jive Software, has open sourced an XMPP server called Jive Messenger. You should check it out -- there is a ton of work on it going on right now.
Patrick
Posted by Patrick Lightbody at Thu Dec 2 06:53:28 2004
Posted by Pingback from Jabber, Zephyr Q&A at Thu Dec 2 08:59:10 2004
http://www.saint-andre.com/blog/2004-12.html#2004-12-02T12:17
My apologies if it it turned out to be a bit of a rant.
Posted by Peter Saint-Andre at Thu Dec 2 11:21:21 2004
Is the client situation for SIMPLE any better than for XMPP?
Posted by Ted Leung at Thu Dec 2 23:19:30 2004
Perhaps the JSF is misnamed, though. Perhaps it should be the Jabber Protocol Foundation, since building software is not its central purpose. I want you guys to succeed, but I'm personally out of resources to help. Between OSAF and the ASF, there aren't any more cycles.
The way you are going to get me to switch is to win on features. And that means somewhere, someone is going to have to build a client and server that do that. Then again, I'm just one guy with a blog.
Posted by Ted Leung at Thu Dec 2 23:32:46 2004
In the higher education world, there are those who would love to get richer SIP clients developed that supported SIMPLE, and they may be able to throw some money at it to make it happen.
Take care.
Posted by John P. Speno at Fri Dec 3 12:55:25 2004
It has tons of features that are not available in many other IM clients on the Palm OS platform.
It also has a low memory footprint. Thanks to the simplicity of the XMPP/Jabber protocol.
Also check out some articles at http://www.chatopus.com/articles on some cool stuffs you can do with Jabber, all from a user perspective.
Posted by Tony Cheung at Sat Dec 4 03:15:35 2004
Jabber is significant because it is a standard.
It is significant because it is a bandwagon.
And I would say: If you can make use of it, please hop on.
Any fool can make a messaging system. But making a system that is a standard that lots of people work together on- that's a big deal.
If you want to do messaging, and you want to do it in the Free world, then you use Jabber.
I think it's a lot like XML. Any fool can make a tree-shaped markup system. That's basic. But making a standard? That's a lot of work. For nothing? No, to make a standard. A standard that does very little? Yes, a standard that does very little. And it's been this HUGE step forward.
At least, that's how I understand all this.
Posted by Lion Kimbro at Sat Dec 4 07:34:35 2004
If someone defines a standard, and nobody implements it. Does it matter? If nobody uses the implementations, does it matteer?
Standards are not automatically good.
Posted by Ted Leung at Sat Dec 4 11:14:19 2004
I understand that there are JEP's that havn't been implemented yet. But, that doesn't really bother me.
I don't see it as a major sticking problem on Jabber.
Posted by Lion Kimbro at Sat Dec 4 16:06:31 2004
I posted the Chatopus, as an example to show some cool uses of Jabber from a user perspective. You might not have a Palm, but there are millions of Palm users out there and mobile messanging is a hot are, too.
Posted by Tony Cheung at Sun Dec 5 03:03:51 2004
Look at motime as an example of integration of a blogging community with a jabber based instant messaging system (with post notifications, too).
Posted by Francesco Delfino at Thu Dec 23 07:32:18 2004
Some of the things you're talking about are also problems in the JXTA community. JXTA has Sun behind it, and paid developers working on a Java reference implementation, but it's still buggy, ugly, and mostly attractive to academics.
One of the great strengths of Open Source software is decentralization and no chain of command, but it doesn't seem to work very well when you are trying to organize people. Decentralized organization is an oxymoron. And if you want to create a unified production effort without duplication of effort, based on a protocol (instead of a working platform, like the Linux kernel) then a conflict between centralized project and distributed effort is inevitable.
The only way I know to help remedy this is to get together a critical mass of interested people and start a social organization that has sufficiently strong shared beliefs to overcome the disagreements: external pressure must exceed internal pressure, or you get an explosion. Structurally, that means having one and only one central "physical" location for all community efforts -- the apex of the pyramid -- and that should be jabber.org. And all the other crap should be destroyed and dismantled. The more bad links, like JabNews, that Google returns, the worse the project looks.
Further, all this posting on individual websites should be contained in a semi-private space where only interested parties will find it. You're airing dirty laundry in public, which is not smart if you want to increase optimism.
These sorts of things are automatic in a corporate environment, but it seems that Open Source projects, especially platform and protocol-based projects, just don't want to embrace them. And so they flounder.
Posted by Brent Gulanowski at Thu Jan 20 12:46:21 2005
To insert a URI, just type it -- no need to write an anchor tag.
Allowable html tags are:
<a href>
, <em>
, <i>
, <b>
, <blockquote>
, <br/>
, <p>
, <code>
, <pre>
, <cite>
, <sub>
and <sup>
.You can also use some Wiki style:
URI => [uri title]
<em> => _emphasized text_
<b> => *bold text*
Ordered list => consecutive lines starting spaces and an asterisk