SixApart aggregation style
One could argue that the single most important feature of Radio – is it’s integrated news aggregator – which could never be built into a hosted blogging tool. I seem to remember Ev speculating on how that could be done. Those of us – who watch these mentions carefully – knew that Pyra was heading towards a “big daddy” scenario – and I’ll predict Blogger will offer some sort of centralized aggregation service – soon.
Meanwhile the king of “DIY servers” – SixApart and their Moveable Type platform – are now about to encroach on Pyra’s territory.
Very intriguing that Marc should talk about aggregating, Blogger and SixApart in the same breath. As Blogger perhaps, sometime, maybe jumps in with a server-side aggregator … the SixApart style, from distributed, small community, server installs to full hosting .. presents a model of how aggregation could grow into the mainstream.
The most simple model of the weblog ecosystem needs publishers and subscribers. Yet I’m certain that the total number of weblog publishers greatly exceeds the number of weblog subscribers.
When you look at the tool process, it makes sense. Blogger sets the standard for easy, server-side publishing. And while there are many great kits for aggregating, the process of finding all the great feeds via directories, or meandering looking for little orange boxes and <link>s, is a pain in the ass for even the regulars.
Beyond the problem of finding feeds, how many internet users are even going to download client software? Even how many of the relatively savvy blogger users? Forget it, they expect it to be on the web or pre-installed (hello MS, hello Apple, or even hello Macromedia).
Server-side aggregation is a hard problem. The number of unique subscriptions will be of the same order of magnitude as the number of subscribers. my.Userland couldn’t manage that much flow, neither could my.Netscape, others tap off at under 10,000. What happens when every taqueria, dance club, and church youth group has a feed?
Oh right, there is a company which handles tremendous crawling volume. But Google is very far from offering any kind of user-oriented, personalized service. One of the most essential features of an aggregator, is the ability to mark/delete stories when read. The volume of that type of user interaction is presently, I think, beyond scalability for server-side app.
Start somewhere in between. MT is an installed server-side app, potentially serving a community of users. A similar service for aggregation would scale well, each install serving feeds for a small to medium sized company, school etc., still small enough to offer all the features of client-side aggregation.
I wish I could start building this, but without time; not too mention the huge unsolved problem of finding rss feeds .. I’m letting the idea loose.