Based on Dru’s 3 Proposals for Open Publishing and Maff’s Audience Hilights the idea behind filters is to create a cross-site, cross-content service, where people can choose to hilight, note, categorize, or other wise express some sort of affinity with content from the IMC Network.

As more people starting using the system, a rich set of relationships, and ideas about the published content will form, making it easier to sort items of interest from the mass of content, and being building a network of credibility.

This is the project I’m really excited about, but I’m moving slow on it.

A medium sized project, larger then the rest of these todos, I’m very cogent of how easy it is to do poorly. I’m struggling with my desire to build a whizbang tool that can be used for translations, and proposed editorial revision, on top of the basic filtering, but I know that I need to build something simple and useable.

Thankfully prior art exists. In many ways, a filter is a blog by another name, and there is a body of knowledge there as the blog community becomes compulsively self-aware. Also very similiar to Phil Agre’s webfilter project, which seems to have died on the vine.

Some issues I’m still grappling with are:

  • Scalability – I’ve gotten the heart of the a mod_perl app for this written, and the basic acts of adding an item to a filter, validating a user, creating a new filter are relatively simple, and light-weight. But I’ll still worried about this intensely database driven app I have spec’ed out in my head.
  • Control of content – whether the focus on IMC is socially or technically enforced. I’m leaning towards saying, its open, and letting the social regulate. Last thing I want to do is get into an arms race with a script kiddie.
  • Some point I’ll need to get a web designer, and a couple of non-techie media producers in on this conversation.