Blog

The BDD Workshop I held at RailsWayCon in Berlin went quite well. I bombarded the 30 or so people with a lot of new and unfamiliar technology - not all of which is RSpec or Cucumber related. Using a Twitter App as an example, we cucumbered, speced, stubbed and mocked around for a whole day and in the end, we had a simple but working Twitter application running. On the side, we talked about the Pomodori method, HAML, used the great RubyMine IDE (I might end up using it for all my Rails development). There were a lot of things I couldn’t get into, maybe next time. ...

Only a few more days, and the Rails World will converge on Berlin for RailsWayCon, an alternative conference to the cancelled RailsConf Europe. I will be teaching a workshop on BDD on monday, 25.5 - including working with Cucumber. See the complete Schedule. I’m looking forward to seeing a lot of good presentations and meeting old and new friends. The lineup of speakers is impressive: Ola Bini, Steven Bristol, Neal Ford, Michael (Koz) Koziarski, Yehuda Katz, Jan Krutisch, Mathias Meyer, Maik Schmidt, Stefan Tilkov, Jonathan Weiss, Heiko Webers and others. ...

Well this is my first post since joining Jens-Christian and Daniela here at InVisible, but the topic is an old one. About a year ago I released the Paperclipped Assets manager for Radiant, which started rather slowly, but has recently taken off and become quite popular. Thanks to the magic of github, I have received numerous patches and improvements and it has begun to take shape. With last week’s announcement of pricing by Heroku and their recent blog about installing Radiant, I decided to give it a try. And, much to my surprise it works really well. Getting used to the read-only file system is certainly something new, but it is surprisely easy and very fast. I love being able to push and pull the database at will, which for small sites in Radiant is a real must to keep everything in sync. ...

twu.li starts to grow. We track almost 8000 twitterers for their updates and collect the URLs. This takes more and more time (there is quite some potential for optimizing it by requesting things in parallel). Right now, that’s about 3 hours to do a full scan (and it will get worse). New URLs can therfore show up with quite a delay. To alleviate that, we implemented a simple priority queue system. The result? The more URLs someone tweets, the more often we will read your timeline. Tweeting less drops the priority for this user and we will use longer intervals. So twitterers like timoreilly will get scanned more often than the huge bulk, that doesn’t tweet as often. ...

I haven’t written the last week, and I guess the suspense is killing all 3 of you, that read this blog. On day 4 (the day before [Cruise](https://twur.li>twur.li was going to be written) I set up the continuous integration and deployment server. My goal was to use <a href=) instead of cc.rb, but after much fiddling with XML configuration files and Rake tasks, I gave up and returned to cc.rb. It works, is relatively easy to set up and suits our needs fine. Why bother? ...

Today was spent setting up infrastructure. Part of this sprint is to learn new things, so we took the opportunity to play around with some of the new features of Rails 2.3. The best thing so far is the templating system, that allows you to define which gems, plugins and other files for your rails app. See below for the template we developed. It’s based on the Rails Template Post by Pratik Naik. ...