The Quest, Revisited

This is the time of year when I cut down on my RSS feed habit, mostly because it all seems to be exactly the same news over and over. I've yet to attempt the RSS-as-email approach again, despite the fact that Ricardo Reyes (who created rss2pop3) was kind enough to send me a brand new version of his newspipe script, which sends the feed items via SMTP (which is much nicer for procmail geeks like me, and saves me the non-existing time to keep twidding ).

(Since Ricardo is currently without a web presence, I'll be putting his source code up on my CVS section (I've yet to unpack the latest version, test it a bit and add a README and the GPL to the archive). The fact that it's taken me nearly three weeks to get around to remember doing this again is a testament to my lack of time.)

This means, of course, that my original is still unfulfilled on the RSS front. My Wiki is the most useful thing I ever put together, I've tamed all my e-mail accounts, have a consistent strategy for managing all my media (and backups), and committed all my important contacts to Address Book while I twiddle (if you're not on Address Book, you don't get synced to my phone, period).

That means I should get back to taming the Hydra (and doing a bit more -centric stuff with the resulting free time wouldn't hurt, either). Here's the plan:

  • Cut down on feeds
  • Get newspipe running, test it against .
  • Figure out if I want a mail-based or web-based aggregator once and for all
    • If I happen to pick web, there is a lot to change on my current Feed On Feeds setup:
      • Do some serious restructuring of Feed On Feeds' innards (data and presentation are jumbled together)
      • Add Bayesian classification (and training)
      • implement XSLT-based views for all the WAP phones I have the (dis)pleasure of testing.
    • If I happen to pick mail, I have to figure out how to rate and/or filter the RSS "messages".
  • Publish the results under the GPL or similar

Either way, it's bound to take quite a while. I must confess the web-based approach has a bit more appeal to me, since it would let me write components I could use for LiveWiki (which would be a neat thing to build in the long run).

This page is referenced in: