A few years ago, highly citational, link/youtube video blogs came back into style again. This time rather than calling them blogs, we called them "tumblelogs." I never really got into it, though I tried, and even my original inspiration for starting tychoish.com was to do a more "tumblelog-esque" blog.It never quite worked out. Then I read this post by Michael Coté which inspired a few things:

First, it got the following title to stick in my head and refuse to get out. Second, it left me with the idea that, although successful, sites like tumblr and to a different extent posterous basically ruined the tumblelogging revival.

Here's the thing about tumblelogs: they worked and worked so well because they were efficient, because the people creating tumblogs were doing something unique and had unique voices, because you could keep your pulse on most of a single discourse by watching only a few sites/rss feeds. And then it became, very suddenly, trivial to make a tumble log. And so everyone had a tumblelog, and it was like blogging was new again, except things "meme'd out" at an epic pace and it became difficult to track what anyone was saying. It was like a distributed denial of service attack on our attention spans.

And as the dust settled, tumblelogs, at least as far as I could see, became less about a sort of delightful amalgamation of interesting content and more about templates, about piping in a fire hose of content from delicious/twitter/etc. So not only were there too many tumblelogs, but the style had devolved somewhat into this weird unedited, awkwardly template-ed mass of "crap" that is (in my opinion) quite hard to read or derive value from.

What Made Tumblelogs Work Originally

  • The systems that powered them were kludgy but they made it very possible to post content easily. That's a good thing.
  • They used a unique sort of web design where design elements (tables/grids/CSS magic,) reflected and accented the content type.
  • They were largely editorial functions. People followed tumblelogs because their authors were able to filter though content with exceptional speed and grace, and in the process of filtering provide real value.
  • They were multimedia, and incorporated many different kinds of content. Not just links, not just embedded youtube videos, but snippets of IM and IRC conversations, song lyrics, pictures from flickr, and so forth.
  • projectionist one of the first and best, was a group effort: when group blogs work, they really work. The tumblelog, seems like an ideal platform for group blogging.

How We Can Make Tumblelogs Work Again

  • We use publishing systems and tools that are unique and that stretch and bend the form. A tumblelog theme for Wordpress, will probably always reek like wordpress. Same with other popular content management systems. Tumblelogs work because they're not just blogs, they need to distinguish themselves both visually, and in terms of how their authors write the content.
  • We undertake tumblogs as a collaborative effort. Group projects complicate things, of course, but they also create great possibilities.
  • Vary content, intentionally, post quotes, chat excerpts, links, videos, lyrics, etc. Make sure that there's a great deal of diversity of content. This is perhaps a problem to be solved in software, at least in port.
  • Emphasize and cultivate editorial voice, and create an interface that forces authors and editors to touch the data.

Thoughts? Suggestions?

ETA: I've started to work on this wiki page outlining a "tumble manager" tool. I also did a bit of textual refactoring on February 27, 2010