Thinking About Testing - Jim Weirich’s Influence

Jim’s discussion was a perfect topic to kick the day off with. I like the idea of test driven design rather then test driven development. I am becoming a believer in how Dave Astell’s views program specifications. I like the idea of rSpec, of creating a specification for your program. Jim tied these together today briefly and then showed some prototype code he’s working called DBC, Design-by-Contract. He threw out the disclaimer that it is merely experimental code. I like what he is thinking about and what he is trying to experiment with. I would like to take the concepts of TDD, BDD and DBC and apply them to actually testing complex business logic. Simple behavior and logic is easily testable with Test::Unit, rSpec and even DBC. Perhaps I should rephrase that and say that the examples out there for these testing methodologies test simple things. Perhaps they should since the examples are to get across how the testing frameworks work. Maybe they are adequate for testing complex business logic problem domains, but there doesn’t seem to be anyone doing those, or at least sharing them. Jim gave good pointers and definitely put some food for thought in my brain. I just need to figure out how to apply it to more complex things, like business rules.

Thinking About Capistrano - Mike Clark

Mike Clark spoke on Capistrano. He gave uber pluggage to Jamis Buck for creating it. I give uber pluggage to Jamis Buck for creating it as well. Great job JAMIS and THANK YOU!! I have used capistrano in the past, and since my production environment can’t see my repositories I can’t really use it in it’s current state. Although I would love to. I could do some tricky things to get it to work, but I have been too lazy to do that. Although after Mike’s presentation today I think I need to quit being lazy and start using it. Before Capistrano came out (well before Switchtower) I had written a suite of deployment scripts. It turned out to be a complete waste of time. It has since been rewritten completely once, and then now I just don’t use them. There are a set of procedures that I routinely follow, but the effort of putting Capistrano to use would definitely make things more manageable in the long term.

Thinking About Linux File Systems - Val Henson</h2

Some interesting pointers from Val.

• If you’re on a laptop, read through /etc/laptop-mode/laptop-mode.conf . You may be able to tweak settings to increase the longevity of your battery life. This is because it is configurable when/how your laptop writes to the hard disk.
• ext2 is super fast for read-only file systems
• ext3 is great for reliability, recovery and generic desktop use
• XFS is great for when you have millions of small files or thousands upon thousands of entries in a given directory.
• ReiserFS 3 is good, but ext3 is better. Supposively ext3 has a more active development and support community, where ReiserFS 3 doesn’t have as good as a support structure. In regards to directory entries, I learned a while back that by default ext3 has a limitation of 32,000 inodes per directory entry. You can change this when you compile your kernel, but it is not recommended because you lose read performance as the number of entries go up.

Open Source QA With Selenium - Patrick Lightbody

I briefly saw an example of how Selenium works at a local Michigan Ruby Users Group meeting, but this did it. Selenium is incredibly cool. The Selenium IDE is incredibly cool as well. I want to dive into this some more and post on this later.

Thinking About ALE -

I didn’t catch the name of the presentor on this, but he works for Zimbra. This was incredibly simple, but incredibly powerful.