Tag Archives: facebook

Novemberween resolution: take (at least) a couple of months vacation from social networking.

I’m moving my online activity here for public communication and (e)mail for private. Ideally, I’ll spend my extra time writing real letters. If you would like to get in touch in a way you can touch, write me:

Haapaniemenkatu 12 A 40
Helsinki 00530 Finland

no accounting for algorithmic taste

I went to check out G+’s new layout, and saw this poorly-done meme image at the top of my list.

Now, I understand the appeal of things like this. They tickle the recognition part of the brain, and since many children play this game, many people will have that recognition and click the Skinner button. Lately, image memes have been spreading in Facebook as well, but at least there you can connect the idea to a person in your social circle. With this one, 134 random G+ Herpderps +1ed it, and that is enough for the algorithm to consider it “hot.” If this is what’s hot on G+, count me out.

On the web, curation is important. If I feel like wasting some time getting exposed to random ideas, I’d like them to have some redeeming social value*.

Boingboing is curated by a handful of people that make it their job to post interesting stuff. Reddit has designed a curation algorithm that works with subcommunity curators that seems to work better than Google’s (at least this crappy image would never have made it to Reddit’s frontpage).

Youtube’s frontpage is a wasteland. There is plenty of good content in YouTube, but with 60 hours of video uploaded every minute, most of it is bound to be crap. In this random screenshot it looks like two clips are recorded by pointing a camera at a TV. Couldn’t their algorithm at least weed those out?

Compare YouTube’s frontpage video selection to Vimeo’s:

I would be interested in 5/6 of the videos on Vimeo. I would avoid 6/6 of YouTube’s top videos like the plague, unless morbid curiosity compelled me click one of them, and then I would feel bad about the decision.

Most of the videos on Vimeo’s homepage are also in YouTube, so there is plenty of good content, but I almost never find it from YouTube itself. It is always a link from a blog or another service. This decentralized curation isn’t a bad thing, but within the site there is obviously a lot of work to be done.

* Thanks Dad.

Sociograph 2.0

An idea that has been brewing for some time…


The 1.0 version was clunky, in Java, and relied on me racking my brain to figure out how everybody I knew was connected and entering the data manually. Day after day I would add to the dataset, until I finally realized that I was becoming obsessive and I had to stop, as the job was neverending. If we were even acquainted in 2002 you are probably in there…


But now it is 2008 and we have fun things like APIs to play with. The only data that I can get to easily to connect people is groups, but I would like this to have more connections graphed. The connections are shallower than my hand-coded graph, but infinitely broader.

I need 5 people to add it in order to release it into the wild, so if you think it is at all neat please do. All suggestions welcome, of course.