Adaptive Aggregators

Album Cover: Wincing the Night Away

"It's like I'm perched on the handlebars of a blind man's bike."
The Shins / Spilt Needles

Posted on April 22, 2006 12:56 PM in Blogging
Warning: This blog entry was written two or more years ago. Therefore, it may contain broken links, out-dated or misleading content, or information that is just plain wrong. Please read on with caution.

My head is so full of ideas right now that they are starting to spill out of my ears. The one I'm about to present here I know I won't have time to prototype or implement, so why not let it free?

One of the most frustrating things about logging in to my Bloglines account after a long, busy day is to see the number of unread posts in the 400 to 500 range. You might be wondering, "why don't you just unsubscribe from some feeds, then?" Because these are feeds I care about. A majority of the time, the number gets really high because of feeds from Slashdot and Digg that have had a few days to build up. When the posts in those feeds get into the hundreds, clicking on them can be quite a commitment. For instance, when I click on the Digg feed, I know I've just put myself on the hook for a good fifteen minutes of reading (at least).

So how can this problem be solved and still allow me to stay subscribed to the feeds that are important to me? Through adaptation.

One of the coolest things about the popular email client, Mozilla Thunderbird, is that it adapts to the way you read your email and removes the clutter so you don't have to. Just because we're talking about spam in one realm and uninteresting content in another doesn't mean the same rules of adaptation can't apply.

Let's say Bloglines allowed me to give each post I read an optional "thumbs up" or "thumbs down" (similarly to how user comments are now handled over at Digg). That would teach my feed reader my reading habits over time. When these habits are better understood, it wouldn't be unreasonable to automatically hide posts from the feed that probably aren't of interest to me, or at least make them less visible so skimming through a long list of posts would become more manageable.

You might think this type of approach would only work for the types of feeds I've mentioned that get a lot of throughput. However, you have to remember that every new post is a potential item coming through your aggregator. In the grand scheme of things, even if your aggregator knows that a once-a-month kind of post isn't up your alley, it has saved you time in the long run. Granted, I know that all of you who subscribe to my feed are interested in everything I say, but I bet there are other feeds out there that aren't so amazingly interesting all the time ;)

I did a little research on adaptive aggregators before posting this idea, and I didn't really find anything. The closest thing I found was over at Geeking with Greg – an article that talks about an actual feed that is adaptive, as opposed to a feed reader. If anyone else knows of aggregators that currently do this (or plan to), I'd love to hear about it.

Adaptation: it isn't just for spam anymore.

Comments

Greg Linden on April 23, 2006 at 8:34 AM:

Hi, Bernie. Findory (http://findory.com) does something like this.

When you use Findory, the site learns from the articles you read, and recommends other interesting articles based on what you appear to like.

Findory does have a feed reader called Findory Favorites at http://findory.com/s/.

If you use the Findory feed reader, you can track specific feeds, but also get a selection of Top Stories from My Favorites, a list of recommended, recent articles pulled from your feeds, much like what you describe.

Findory can import OPML or your subscription list from Bloglines if you want to try it out.

Findory does not do explicit ratings -- thumbs up or thumbs down -- because most people won't bother entering that kind of data. Findory instead uses the implicit information about your interests it gets from paying attention to the articles you read on Findory.

Permalink

Bernie Zimmermann on April 23, 2006 at 12:49 PM:

Thanks for the information, Greg. I don't debate Findory's usefulness, because I believe it takes quite a while to find a list of feeds and content that suit your interests. The problem is, I've finally reached that point, and rather than trying to find more content that is similar, I'm trying to take the feeds I'm currently subscribed to and focus them down to the subcontent that is important to me.

I remember that Asa Dotzler ran into a problem several months ago where people were complaining that they didn't like it when he blogged about cats or space, and were only interested in Firefox related content. Readers were trying to put the onus on the blogger to weed out the content they didn't like. I personally think this is backwards. Through the ability to subscribe to specific categories, or better yet, the ability to rely on one's aggregator to only allow the interesting content (specific to the reader, of course) to get through, I think this type of situation can be avoided, and everyone's happy.

Again, though, thanks for providing more information about Findory. I am sure many people will find a service like that most useful.

Permalink

Greg Linden on April 23, 2006 at 1:59 PM:

Thanks, Bernie. Just a quick followup.

Findory Favorites (http://findory.com/s) does do something very close to what you want, taking the feeds you are currently subscribed to and focusing you on the most important posts of the day. The recommended articles in the Top Stories from My Favorites are exclusively articles from the feeds you list as Findory Favorites.

Might try it out if you haven't already!

Permalink

Post Comments

If you feel like commenting on the above item, use the form below. Your email address will be used for personal contact reasons only, and will not be shown on this website.

Name:

Email Address:

Website:

Comments:

Check this box if you hate spam.