Knowing my passion for RSS, James sent over this post from Marshall Kirkpatrick yesterday. Here, Marshall talks about “management by feed”, which I take to mean “using RSS to get the right information to the right people.”
And if that’s not what he meant, well, then let’s just say whatever it was he did mean kicked off a long train of thought. Instead of dumping out another novel, I’ll break this into separate posts: feed filtering, social feeds, and feeds as protocols.
Smart Feed Filtering
Every morning I sit down in front of a large number of feeds, and a still larger number of unread posts. I’m continually trying to come up with schemes where I can have “small victories” over reading the problem of finding the good content in that big mess.
So far, all of the tactics I’ve come up with are “manual,” in that I have to do all the filtering myself, perhaps arranging feeds in different folders (like “!mustread” or “!readonfriday”). My solutions so far haven’t even approached the helpfulness of the best email tip ever: keep your inbox empty, act and archive. I may use the nuclear option some day soon: Mark All Read ;>
Below are some filtering based tactics that I think would win me enough small victories to get my feeds under control, to manage my feeds as it were, and better manage my information life by feed. I suspect that none of these ideas are new and that, if I searched around, I’d find other people who’ve documented them. But, what the hell, repetition is the surest path the success. Many of these ideas are not strictly for behind the firewall.
By “Smart Feed” I essentially mean filtering. If you’re familiar with iTunes, then you probably know about iTune’s Smart Playlists. You can create playlists based on searches like “all songs not played in the last 5 days that are rated 3+.” The idea with that one being that you have a list of music you like that you haven’t heard recently.
Here are some “Smart Feed” filtering ideas:
- Show only feed items that are popular URLs pulled from digg, del.icio.us, other “perspectives” (slashdot, Google News, memorandom, etc.). This type of Smart Feed would match the URLs for posts at ranking sites, allowing you to tell your feed reader “show me all items/URLs in my feeds that are digg‘ed at X level of popularity.” The point of this is that you can read through all the popular links in your feeds. Also, this double filtering: you’ve already filtered by the authors you want by picking feeds. The hope is that with these two layers of filters (narrowing by authors, and then by what other people think is popular in that sub-set) you’ll read more good stuff instead of uninteresting (to you) posts, either in your aggregator or at those ranking sites.
- Show only posts in my feeds that have been tagged with some set (including just one) of tags. I would tell my aggregator that I’m interested in a set of tags (and perhaps 1-2 order of related tags to help me discover new things), and I’d see only posts that were tagged appropriatly. Here, your aggregator would integrate with the likes of technorati and del.icio.us to find out what tags a URL had. FeedLounge has tags, but I suspect it doesn’t have as many tags as technorati, del.icio.us, and itself combined.
- Show only posts that my trusted group of friends, or other groups o users, have read. For example, what items in my feeds have the sales people been reading? Or, what items in my feeds have my co-workers read? This is a great passive filter as it doesn’t require additional input from users (unlike ratings and tags). The idea here is that I might want to prioritize my reading by catching up with what my friends have read. Of course, getting the inverse would be handy too.
- Building on the social network aspects, an aggregator could also show only posts that my trusted group of friends had rated high. It would also be interesting to know what other groups had rated high, for example, management. The data could be made anonymous at the group level: all people at such and such level in an organization really liked this article on off-shoring. Get ready… Of course, preventing that kind of transparency is often the key block to all this type of thinking. We’ll get into more of this in the second part of this little series.
The key to most of these are that they’re not feeds themselves, their views/perspectives of the feeds I already have. They’re filters that can be quickly turned on or off, or mashedup together. The static nature of a feed to accomplish the above seems to preclude flexibility and would leave me with that feeling like I was missing something: even if I never turned the filtering off, at least I could.
Scaling
The main concern with the above is the amount of real-time data required to filter the feeds. Pulling all that data from different sources in realtime is challenging, both on the client and server side. On the client side, I wonder if doing most of that filtering in the browser would help: you could even use Greasemonkey to do it. On the server side, I suspect caching much of the data as static instead of calculated at runtime would help…if they’re not doing that already.
(Management by Feed/Feed Management Series: part 1, part 2, part 3.)
Recent Comments