Regarding Dave Winer proposal to create a share your opml app, I’ve put up a test site at:

http://dev.glistn.com

Just testing functionality. This is a development site!

1. Join

2. Upload OPML file

3. Edit permissions of the individual feeds in the OPML by clicking on View files and then on [edit] next to the file you want to change. Submit it.

4. Now your permission based OPML will be available (only after you have edited permissions at least once)

5. View the OPML and share it at http://my.glistn.com/username/filename/

This file is hosted at Amazon s3.

After seeing GestureBank at work firsthand, I must believe that any open source news project like this must use implicitly gathered data and information as well as OPML uploads, Digg type voting, and profile creation.

Therefore, we need a system to collect behavioral statistics from search and surf.

The trick is to be able to use this information anonymously while also being able to couple it with the explicit information that a user gives publicly.

We are close to getting some code up. Have patience, thanks.

We need niche content that is personalized, not just personalized mass media.

A reading list give us the general idea of interesting people on given topics.

We need to extract those relevant posts, plus others into niche categories of threads.

Links from posts aggregate threads.

Page Rank determines more important posts.

Past behavior determines another level of likely relevance

Past behavior tells if a portion of the link was read already, therefore, don’t show it.

Past behavior can be told by anonymous pools of metadata, attention recorders, cookies from contributing blogs, glistn click behavior

Where does alex barnett belong microsoft or opml category. . .both of course

So organization happens at the post level even though community or social networking happens at the aggregate feed or reading list level.

Thanks for stopping by. Lots of work, but I plan to have something live rather quickly. I’ve come to believe that any truly important filter of news, information, or data we consume will have to be open source, or else it can’t be trusted.

Like we will need to own our Attention data, we will need to own, understand and control how we process incoming data, and so this project is born.

Glistn will attempt to empower you to filter and consume the web, but not try to control your lens.

Off the top of my head, here are a couple things I promise.

1. Any data collected by the service will be available for you to export in at least one open format. It’s yours.
2. It will be open source, so you will be able to see exactly how it works and change it if you like. Or run the software yourself.
3. The system will be built on top of APIs from the onset. This way every single possible action can be called remotely. We are not looking for eyeballs here. That would change the way we designed the product from the start and have its own consequences. We don’t need that baggage.

4. It will have a plugin architecture to accommodate all the developers who are more creative and faster than we could ever be.

More to come. Send ideas, or even loud criticisms.

If you are subscribed here you’ll hear about the project as it makes progress. or signup at the site to be notified of the launch.

Matt Terenzio’s general blog about the web is available at http://everybuddy.org!

Ping.