|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
Philosophy of aggregator.root  A few days ago, after running a survey that showed substantial interest among Frontier developers in aggegator.root, the engine behind My.UserLand, I decided to go ahead and release the code. It installs reasonably easily, but not quite as easily as it could, if we knew more about how it will be used. Our needs are special because we're aggregating a lot of channels. How should it be set up on your machines? Let's figure that out. And on what terms do you want to receive it? One of the reasons I wanted to study open source is so that we could do it more formally, in a way that makes sense for UserLand and the community of developers that use Frontier. I wrote a bunch more about this on the home page of the new aggregator weblog. The code of aggregator.root OK, you want me to show you the code and shut up! Here it is.  At this stage, I'd only recommend it to experienced Frontier developers. You must have a full installation, including Manila and mainResponder (standard stuff in 6.2.1). If you're not a pioneer type, sit this out for a bit. And who knows how updates will be distributed? I don't. Lots of questions to answer, but if you'be been itching to play with syndicated XML content, go diggin, but remember to do a backup. What changes? It's been a while since I've written about syndication. I'm going to re-read this piece and share some comments. If there are more aggregators, that might change a few things. A quote: "My company, UserLand, has been working with Netscape and others on a reversal of the Vignette philosophy. Instead of being a flow concentrator, we propose to be flow distributors, with value flowing in the opposite direction, from the source of the content to the source of the click." That's a key point. By releasing aggregator.root, we're not acting solely as a flow distributor, we're a technology vendor. If this release takes root (pun intended) then we are providing technology to people who are flow distributors. At the same time we run our own flow distributor. (And we also push content through the network, Scripting News and other UserLand sites are channels.) Morning stuff Brent reviews Mac OS X. Salon: Who wants to date a dot-com CEO? Don't forget to file your Behind the Curtain photo galleries. (Don't you hate it when some people do their homework early and then remind you when it's due? )A Manila site served on Mac OS X, in the Classic environment. Rick Winfield wants simple XML. "I had gotten caught up in the 'coolness' of my new system and the XML I could make it produce without considering what would best serve my users." I just re-subscribed to the CMS List run by Phil Suh and Cameron Barrett. Somehow I was logged off. Phil asked a question about CMS client apps, they use Frontier at Phil's company, I just posted a pointer to Radio UserLand's feature list. It's squarely in the category he describes. Yesterday's discussion on the Decentralization list got to some interesting places. I'm still not sure what Mithral's innovation in P2P is, some people are pretty breathless, I want to know why. Clay Shirky is a believer in P2P as a way of harnessing the compute power of a network of PCs. I'm skeptical, I wish I could see an application that isn't about human intelligence. Sometimes when posting on a mail list, I write something that should be on Scripting News. Intel has been trying to hold a meeting of a P2P Working Group. I signed up, of course. So did a lot of other people. It's growing like a virus. "Watching Intel up the size of the meeting room twice, I wonder where it will stop. How many people like me want to be at the summit? Maybe they'll need to rent PacBell Park?? Maybe they'll have to forget about the meeting and use the Internet to figure out what P2P is. Wouldn't it be funny if you needed P2P to figure out what P2P is??" I'm beginning to think that P2P is exhausted. I know it makes me feel exhausted, it seems like a chance for us to drift again, away from the serious work we've been delaying in building the decentralized Internet at a content level. There's a philosophical face-off. People who view the Internet as stages with audiences (the TV model) and people who see it as something new that's not like TV. I'm in the latter group. So far the money has been betting on the former group. I'd like everyone who writes to have a website, and to have tools for finding the stuff they want to connect with. So far the Internet "business model" has been an exercise in capturing bits behind specific domain names. Instead, I think the technology industry should get back to technology and let the people write the Internet and stop worrying so much about capturing people. History says you'll be routed around, if not by the Internet, by the grim reaper. Remember the barking farting chihuahua thing. Here's an example of the world going the right way. At first I was not happy that eGroups, now a part of Yahoo, started putting the ads at the top of emails. The ads really disturbed me, made the emails look like trash in my mind. As time went by it got worse. Then the ads started disappearing on some of the lists I belong to. Someone paid the $70 to get rid of them. Now they host a million lists at eGroups. If they all paid the $70, that's $70 million. A nice chunk of change. And now they have no stake in our bits. They're providing a valuable service. So charge for it. Makes sense!! So thanks to Yahoo for shifting the conversation. The new business model for Silicon Valley -- create valuable service through technical excellence and commitment to customers, and stay out of their way and earn a nice profit. BTW, that was the *original* business model for Silicon Valley. We lost our way in the latest euphoria.
|
 
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
© Copyright 1997-2005 Dave Winer. The picture at the top of the page may change from time to time. Previous graphics are archived. Previous/Next |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||
