How to decide in less than 5 seconds whether to keep or kill a feed

I keep noticing how my CPU is getting sucked up to an astonishing degree by NetNewsWire (as I write, it’s doing a check on feeds and the machine is on 100% CPU).

And given that I had about 630 feeds, I thought there might be a connection. Particularly because it means there are periods when I type and nothing happens. NNW will happily use up 80% of CPU at times like that.

So, obvious tactic: kill off some of those feeds. After all, it’s not as if I ever got round to reading all of them. I’d read a few all the time, many a little of the time, and some of them – I’ve no idea how I put them there or why. Must have seemed sensible at the time.

So I’ve been working my way through. And here’s how I decide whether to kill or keep.

Kill if:
* daily content consists of “today’s links:” followed by links and no new content.
* posts consist of no new content that I couldn’t find elsewhere, or just think of myself. (Musings on the crack in their living room wall or the new curtains they bought don’t count.)
* Haven’t updated the blog for more than 30 days (a “dinosaur”, in NNW’s parlance. Though I kept Tim Berners-Lee’s blog.)
* feeds are partial. Listen, don’t try to tempt me to come to your site to read your fantastically insightful things. There are very few organisations that can manage that, and of them, most are newspapers or very high-flown analysts of a sector. I just don’t bother with partial feeds any more. Apart from anything, most people can’t (a) write a good teaser (b) come up with any sort of analysis that makes it worth clicking through. Your overall traffic will surely be higher with full feeds. I’ve made this site full feed from day 1 because I couldn’t see any benefit from making people come to it. Yes, a few sites do deserve a clickthrough. But even then, it’s a barrier to reading. Barriers online don’t help.

Keep if:
* they’re in a sector where I want to hear the authentic voice of the user or programmer or problem solver
* they’re consistently able to throw new light on things I’m interested in
* they have a track record of telling you about things before they get big
* they’ve got a full feed.

Using just that criterion, and devoting about 5 seconds to deciding with each feed, I’ve already whittled it back to 550. I’m going to repeat this a few times. I think 200 is a good intermediate target – but I suspect that the ideal is a lot closer to 100.

Update: now down to 399 feeds – which translates to only 10,098 unread articles. Ho hum. I’m sure I’ll find more to whittle away.


  1. I hover around 90 feeds and a lot of those could probably go too – too many aggregators reaggregating each others content at the moment.

    And as for Guardian guy’s feed……

  2. One other thing you can do to up performance in NNW: in Preferences, under Downloading, change “Concurrent downloads” down to less than five – that often has a positive impact on performance.

  3. Charles

    Friday 15 February 2008 at 5:16 pm

    @L – that’s pretty slim.
    @Ian – I only worry that if I did that at present, it would take longer than 30 minutes to do the update, and then the next one would start before the first had finished, and the universe would implode.