0xDECAFBAD

It's all spinning wheels and self-doubt until the first pot of coffee.

On the destructive/constructive nature of news aggregation

Playing with a range of news aggregators once more, since Radio UserLand is making me itchy again. Pretty much the only ones I really like are Radio, AmphetaDesk, and sometimes Peerkat.

For the last week or so I've been mostly alternating between Radio and AmphetaDesk. The difference in aggregation styles is interesting: While Radio slices and dices and aggregates feeds by items and serves them up to me in an interwoven chronological order, AmphetaDesk serves my feeds up to me whole and in order of feed modification. I'm not sure which I like more now.

I like Radio's style, because I see what's new and only what's new. Usually. But, I like AmphetaDesk's style because I see everything, and have realized that I miss things with the rush of new items from all my feeds. For instance, if someone posts something once per day, I'll likely miss it with Radio unless I check every hour. But, with AmphetaDesk, I get to see what's new with every person or source whenever I check, and I only miss items if that feed has scrolled them off.

Shelley Powers of burningbird.net wrote a bit about the context destroying nature that RSS and aggregators have on weblogs. I agree with her somewhat, in that pulling the words out of the surrounding context of the blog and its presentation and community has an altering effect, I wouldn't say that it destroys the weblog.

For me, when I see something on my news aggregator, it's as if I'm overhearing it from another room. I don't get the whole context or see who else is listening or responding, but I hear the gist of something. And, when subscribed to 100+ RSS feeds, it's like I'm floating in this Nth dimensional space where I can overhear voices from hundreds of rooms without being overwhelmed. When something triggers some of my mental filters and watchwords, I click the link and delve deeper.

There's no way, without an aggregator, that I'd be able to track 100+ people and news sources in a day. But, because I can, I've been able to learn and discover things and hear voices I never would have before.

But, I think the way AmphetaDesk merges these sources might be a nicer alternative. By not chopping up the feeds, some intra-item context is maintained at least, so I can see developing trains of thought.

Okay, must go back to work now.

shortname=oooaai

Archived Comments

  • Rael
    Howdy, If you like (in any way ;-) Peerkat, you might want to take a gander at the dynamic duo of Blosxom (Weblog application) and Blagg (RSS Aggregator). Rael
  • l.m.orchard
    Wow. I've been playing a bit with blosxom, resisting the urge to do any tinkering that would complicate it. It's like a weblog haiku on my iBook . And Blagg looks pretty dang nifty too. Must pla y with these more. It even does the sort of thing I was thinking of with GroupWeblogWithRadioUserLand, only in a terribly less complicated way. :) So the other thing I have to ask: How does one become a researcher for a major publisher and get to play with things like this all day? :)
  • Prof Avery
    I've been putting off installing AmphetaDesk on my OS X machine at home, and now I suspect I won't bother -- Rael hit another one out of the park! Me too, how do I get a job like this? P.S. You might find my XML-RPC interface for Blosxom entertaining.
  • l.m.orchard
    I just went nuts and via a lil regex mangling, threw my whole Radio UserLand subscriptions list at Blagg. Let's see how much I can torture it :)
  • Rael
    > Let's see how much I can torture it :) So nu, l.m., how did it go? > and get to play with things like this all day? :) Make that all evening and you've your answer ;-)
  • Rael
    Fresh from the oven: you might enjoy Blaggplugs, plug-ins for Blagg hooking it into the Blogger API, Movable Type, et al.
  • Rael
    Prof: your BlosXMLRPC is a hoot!
  • l.m.orchard
    Ahh, blagg. Another haiku of code :) Torturing blagg with my pile of subscriptions seems to have gone alright. I need to tweak around with the templates for blosxom (Blog Some?) to make it more pleasant to visually scan. Also need to tweak the max # of items on the page. Maybe some sort of "Next/Prev 20 items" pagination would be nice. Also, templatizing how blagg constructs posts from RSS items would be hot, too. First snag I see though, which you probably already know about, is the way blagg determines whether or not an item is new. Blagg, of course, uses the title. However, some of my feeds' items never have titles (ie. Radio UserLand peeps who haven't turned on titles), and others often have the same titles (ie. RSS feeds from wikis). I've been trying to figure out the best way to do this in general, and all my solutions keep ending up at a bruteforce MD5 of the whole item. Of course, the problem in doing that in blagg is that MD5 doesn't come out of the box with perl in OS X, which kills the whole drag-n-drop installation theme. (Nevermind that MD5ing everything might be a tad slow.) Hrm. Still playing.
  • l.m.orchard
    Oh yeah, and one other blagg thing: I see the Blaggplugs. Keen! Something that would be nice, though, is if the RSS feed source could be separated from the post title and passed as a separate parameter to the post() sub of the plugin. Then I could even do something like transform the source attribute into a category or something, should I make a metaWeblog API plugin. Just supposing I did, that is.
  • Morbus Iff
    > blagg determines whether or not an item is new Hmm. Yeah, MD5 would be a little slow. Perhaps the filename could include a character length count, something like "raelity.617.title". That should cover Wiki changes (as a same count edit would be rare, I'd hope). That'd cover RU a bit too. Speaking of which, I just sent Rael a Blaggplug that pings weblogs.com with an update of your blog. I've got another one planned for either today or tomorrow as well.
  • Rael
    Another about-unique index I was toying with was to select the xth, yth, and zth character of the description and include that in the filename. What are the chances you'd get the same title from the same feed with the same 3 arbitrary characters in the same 3 arbitrary spots? What do you think?
  • Morbus Iff
    Seems suitably random to me.
  • l.m.orchard
    Hmm... You might still be in trouble with a wiki RSS feed like the Meatball wiki RSS feed or the MoinMoin RSS feed where edits are minor and the page is left largely unchanged. But, it sounds like a better step.
  • rael
    New email and IM Blaggplugs -- get 'em while they're hot!
  • rael
    Woohoo! I'm now entirely self-hosted (read: Blosxom and Blagg)!