The Exploding Edge
Editors' Note: The following is this week's issue of SuitWatch, senior editor Doc Searls's biweekly newsletter out business and Linux. Subscribe to SuitWatch here.
I'm writing this the morning after Supernova, Kevin Werbach's new conference on decentralization. In his introduction, Kevin likened much of what's happening in the tech world right now, in spite of economic troubles, to a supernova-type explosion: all kinds of hot stuff happening in the expanding periphery while the center becomes a relatively cold vacuum.
Decentralized stuff scales in ways that centralized stuff can't, he explained, offering the uncontrolled explosion of wireless (WiFi) internet access as a standout example. Consistent with this trend, the conference kindly supplied free WiFi to attendees--or to anyone else who happened to be hanging around the underground meeting area at the Sheraton where the conference was held.
It was interesting to me that my laptop was able to detect two signals: the free one for the conference and the fee-based one for the hotel. Given the limited range of the conference's signal, it surely didn't take any business from the hotel's system (which I'm guessing charged the usual $7 to $10 per day). Just as interesting was the situation at the much cheaper Quality Inn up the street where I was staying. While the room provided the customary combination of rough sheets and Lysol odor, it more than redeemed itself by providing free internet access through the blue Ethernet cable hanging out of the wall by the desk where I'm typing this right now. All of this connectivity--both for free and for fee--demonstrated explosive "scaling" by the Internet itself, which is the most decentralized creation in the history of civilization.
The opening talk at the conference was given by Howard Rheingold (he's a good guy who's been in front of interesting trends for a generation or more), whose new book (the latest of many) is Smart Mobs. Every time power decentralizes, he said, there's an opportunity for innovation. Among his examples were [email protected], teenage cell-phone messaging in Scandinavia and Japan and the Net itself.
In a later talk, David Isenberg showed how it was possible for two PDAs to serve as phones that communicate over the Net, with carrier-grade sound quality, with or without wires, at no cost and with no intermediating telephone company--decentralization in the extreme.
David wrote a famous (or infamous, depending on your point of view) 1997 paper titled "The Rise of the Stupid Network". He was at AT&T at the time, and the paper so insulted his employer that he left shortly thereafter and immediately made a career of pointing out--often to great effect--how stupid great networks need to be.
"Why Stupid is Still Smart" was the title of his Supernova talk, and "Distance is dead" was his opening point. He went on to explain the need to keep making the End-to-End Argument first made by David Reed and others in the landmark 1982 paper by the same name. This argument was so persuasive that it served as the conceptual blueprint for the Net. Yet in spite of its success, the same argument remains opaque to vast populations who aren't hip to the Net's profoundly decentralized nature. This list includes Congress, Hollywood and Microsoft.
Another speaker at the conference was Dan'l Lewin, a Silicon Valley veteran (Apple, NeXT), whose latest job with Microsoft is running .Net business development. Lewin's talk opened with good stuff about the history of the Net but closed with a complex and vague description of the .Net development framework (or whatever it is, which remains unclear). Amidst nice-sounding talk about interoperability and how "it's all XML, right?" Lewin was peppered with questions about licensing and patents. "It was wonderful how you just snuck in stuff about licensing a schema", one audience member said. Why license stuff that's supposed to be Net infrastructure? people wondered. Why wouldn't Microsoft give away at least some of the foundational stuff? And why trust the company in any case?
From this point forward, through subsequent panel discussions (on collaboration, web services, weblogging, "rethinking telecom", etc.) and talks, it became clear that the decentralized Net supports two related developments about which centralized (and centralizing) organizations are still in denial to some degree: 1) commodified platforms and 2) simple and useful APIs, formats and other standards that allow for conversations and relationships among many different kinds of computers and applications. Linux and Apache are in the first category and web services are in the second.
The biggest and best-exposed irony at the show was the distance between the rhetoric and the reality of web services. One panelist talked about the distance between the dream of connecting everything and its manifestation in SOAP and WSDL, which he compared to the Java dream/reality outcome. "We do a lot of work, but there's not a lot of uptake. Google (and other services) are cool, nice toys and fun, but it's not like there's a whole groundswell of apps on the Net."
The audience handed him his head. "You're looking at the wrong developers. Get with the program. It's happening."
In fact, it was happening right there, where the Net met the room. Audience members were reporting the panel, live, on their weblogs. Every one of those blogs notified the world every time an update was saved. Those notifications were picked up by other weblogs, (and, perhaps, if a story became newsworthy enough, by more formal news organizations). Clues proliferated at the speed of packet traffic. Cluelessness was widely exposed, and corrected, just as fast, all thanks to real live web services. Programs were talking to programs, and services to services, through RSS, XML-RPC and open APIs published by the likes of Google and Blogger--all running between a wide mix of platforms and programs.
Worse for that unfortunate panelist, the authors of some of those very services and products were in the room. I'm told one audience member--yours truly--said "we're hacking your ass" to another panel. I don't remember saying that (I recorded it, so I'll have to check on it later); but it was true.
It's an easy and common mistake to understand any topic in terms of big company hype. It's also natural to expect big companies to do all the significant work and not be confined to the heavy lifting that only big companies can do.
IBM, Sun and Microsoft have done a fine job of branding the web services meme. But when real web services get delivered in a decentralized way by non-brand-name companies--or by helpful hackers and other unusual suspects (e.g., Google)--nearly everybody misses the story. At first.
But sooner or later the facts overcome the hype and everything becomes clear--whether it's the Net, Linux, web services or any other new and decentralized development. Thanks to web services, the time that process takes will be explosively reduced.
Doc Searls is senior editor of Linux Journal.
email: [email protected]