return to first page linux journal archive
keywordscontents

Linux for Suits

The Message

AWZ.com is a teen site running on Linux that uses what may be the most popular teen application of our time: instant messaging. Lately, IM has become a hot topic, not so much because it's such a hot application (which it is), but because AOL seems hell-bent on dominating the category. For years, the company has been bundling AIM (AOL Instant Messenger) with not only their own proprietary on-line software, but with every copy of the Netscape browser. AOL also owns ICQ, the popular IM system developed by Mirabalis in Israel a few years back.

According to Tapan Joshi, the president of AWZ.com, AOL is trying to dominate IM the old-fashioned way: by working (both in software and the courts) to keep anyone else's instant-messaging software from interoperating with AOL's, regardless of the inconvenience this causes to AOL's own customers. In an April 27 op-ed piece in the San Jose Mercury News, Joshi lamented it isn't just Microsoft and Yahoo! that AOL wants to keep away from AIM and ICQ. It's little guys like AWZ.com.

Reading the piece brought to mind the last conversation I had with another IM pioneer: Udi Shapiro. It was in early 1998, not long after Udi had bought his company, Ubique, back from AOL. The on-line giant originally acquired Ubique for its technology, which subsumed and transcended ``chat'', ``buddy lists'', ``instant messaging'' and the other stuff AOL was also busy building into a Ubique competitor, the now-hyperfamiliar AIM. When it became clear that Ubique was getting the short shrift at AOL, Udi quietly repurchased the company, searched around for a proper benefactor, and found one in IBM's Lotus Development Corp. LDC purchased Ubique in May 1998 and now sells its technology as Lotus Sametime (which PC Magazine rates ahead of AIM, by the way). Meanwhile, Udi is back where he started, at the Weissman Institute in Israel, working on computational and biochemical theories of the origin of life and stuff like that.

I'm glad I had that one lunch with Udi, back there between AOL and Lotus. During that lunch, Udi said the main lesson he learned at AOL was that instant messaging ``changed the sociology of telephony'' at any company that used it widely. Thanks to the buddy list in AIM, it became bad form at AOL corporate headquarters to bother a co-worker with a trivial call when a trivial instant message would do a better job. ``Got a minute to talk?'' and ``Where are you going for lunch?'' were better asked on the instant messenger than on the phone. New manners developed for both phone and IM. It was a remarkable development, Udi reported. Typically, this realization was mostly irrelevant to AOL's own concept for AIM, which was to serve as a frame for advertising messages delivered to the service's heaviest users: teenage girls chatting about boys and homework.

The next visionary to deliver insights to me about instant messaging was my friend Perry Evans, best known in our industry as the creator of Mapquest and to the rest of the world as the guy whose idea it was to create flavored bite-sized shredded wheat (or something like that), when he was a hungry kid working for Nabisco. Now the president and CEO of Webb.net, Perry let me know over drinks in Denver last year that he and Webb were doing for next-generation instant messaging roughly what Transmeta did for Linux, which was to give its progenitor a day job.

The Linus Torvalds of next-generation instant messaging--and a low-profile employee of Webb.net--is Jeremie Miller, a quiet, modest and brilliant programmer who lives in the farm country of Iowa. Jeremie's Linux is Jabber, ``an open-source, cross-platform, completely extensible, end-all to Instant Messaging as we know it'', or so says the FAQ at the Jabber.org web site. But it's more than that. To understand how much more, you have to know two things.

First, instant messaging as we know it--almost entirely through AIM and ICQ--is a client/mainframe affair. When you're on an AIM client, you're hooked into exactly one server, in Vienna, Virginia. When you're on an ICQ client, you have the same relationship to another singular server. Same with Excite and Yahoo! That's four services, four giant servers and nearly zero interoperability between them--just like it was in the good old days of dial-up bulletin board services and on-line services. To be fair, Microsoft's MSN Messenger and Lotus's Sametime are both client/server systems. You can buy and deploy them on a server-by-server basis. But every one of these systems offers nothing in the way of real Internet Infrastructure, since they're as proprietary as a company boardroom and as open as a brick.

Second, about all most of us know so far about instant messaging are the three features we associate with AIM and ICQ: buddy lists, chat and instant messages. We'd also rather ignore AIM's and ICQ's other agendas, which are to serve as consumer marketing vehicles for branding, advertising or both.

Jabber is a totally different animal. Here's the way Perry put it when he blew my mind back in Denver: ``Jabber puts a living XML document in the middle of a real-time conversation.'' That means the content of an instant dialogue can be much more than typed text. For example, it can be a legal brief, collaboratively edited by a number of lawyers in a firm. It can be a live auction, with bids published for everyone in real time. It can be a repair history for your car, viewed by you and your dealer on your computers while you talk over the phone. It can be, well, just about anything.

On top of that, Jabber is completely XML-based, open source and deployed as an infrastructural service on the Net. An enterprise, a household or an ISP could have a Jabber server, just like they have a domain name server, a web server or a mail-hosting service. What's more, Jabber translates to any familiar instant-messaging flavor, including ICQ, AIM, IRC, MSN and the emerging IETF/IMPP protocol. Everyone can use their own client if they like, or they can use a Jabber client. Want more? Jabber also features message filtering, MIME encoding, crypto (e.g., OpenPGP) and game communication. Webb's Andre Durand, founder of Durand Communications and now general manager of Jabber.com, Inc. (a subsidiary of Webb) describes Jabber as ``a real-time router for XML documents that has the potential to bridge not just proprietary IM networks, but wireless devices, embedded systems and a whole slew of new, real-time messaging-based applications which have yet to be developed.''

``You can create all kinds of applications that run on Jabber,'' Perry says. ``Voice-over IP integration, software development collaboration, community buying aggregation, auction and reverse auction systems, enterprise groupware--the list just goes on and on.''

Right now, Jabber's profile is still low. It's kind of like Mosaic before it was released by Marc Andreessen and his buddies at the University of Illinois. (You can find an early Jabber version deployed at Corel, if you want to start trying it out.) But a groundswell is building. In February, when I was on an open-source panel at the New York New Media Association, Daniel Frye of IBM began gushing about Jabber in a way that put even my own enthusiasm to shame.

So clearly, this is a space to watch. Unfortunately, Mr. Joshi of AWZ.com isn't watching closely enough. In his op-ed piece, he correctly calls AOL's belligerence ``short-sighted'' and instructs the company to ``look no farther than the Linux operating system.'' He predicts ``it is very likely that AOL is going to learn the Linux lesson the hard way.'' But he misses Jabber.

Jabber will teach the lesson, but it won't be hard for anyone. That's because Jabber doesn't compete with AOL or anyone else. It simply enlarges their scope. ``I look at instant messaging as a wild frontier,'' Jeremie says. ``Right now, there is nobody around. We're free to build all kinds of stuff, and everybody else is free to add value to what we're doing.'' Sound familiar? Keep watching, or go to the developer section of Jabber.org and see what you can do for the cause.

Doc Searls (doc@ssc.com) is Senior Editor of Linux Journal, co-author of The Cluetrain Manifesto and a member of the advisory board for Jabber, Inc.