RailsConf Day 2

Carl woke me up early this morning by jumping around on my chest. I got ready and drove back down to Chicago for day two of RailsConf.

The first session of the day for me was Obie Fernandez's Thoughtworks on Rails which was a broad overview of the rails projects that Thoughtworks has done for its customers after introducing it into their development environment. Nothing too technical, but useful to see the lifecycle for a rails app from the point of meeting with the customer to creating "stories" as they put it, to coding individual pieces, to quality assurance testing, to final deployment. I couldn't help but think about how many people are involved in these "normal" development processes versus things at DLS where one developer has to take a request from another staff member and develop, code, test, and deploy an entire app himself.

The next session was Scott Raymond's Lessons from Blinksale & Iconbuffet which I honestly didn't really pay much attention to because I was doing things on my laptop the entire time.

Matt Biddulph's presentation Putting the BBC's Programme Catalogue on Rails was fascinating. A very small app (3985 lines of code if I remember correctly) to manage gigabytes of data spanning many decades. Matt explained how the app was designed and touched on some technical details for fulltext searching with MySQL.

After lunch, I attended Badrinath Janakiraman's Legacy DB Schemas with Ibatis for Ruby presentation which was a great eye-opener for me, as I had no idea there was this better way of presenting crappy old database schemas to Rails without making views or hacking up ActiveRecord. However, his advice was still to try to use ActiveRecord whenever possible because Ibatis is a lot of manual work and somewhat error prone.

Next was Accelerating Rails with Lighty from Jan Kneschke, the author of lighttpd. A good technical presentation on how lighttpd was designed and some of the benefits it has over Apache and other web servers. The X-Sendfile feature described was fairly interesting, though I don't really have any scenarios where I'm serving large files.

After this session, all three of the rooms were combined again into the large area and Mike Clark gave a presentation Testing Rails Applications which drove home the point of writing proper (or in my case, any) unit tests. For some reason I suddenly got really tired at the end of this presentation and could barely stay awake (I blame Carl). After Mike's talk I went out to my car in the parking lot and took a nap, and then went to Chili's for dinner.

Back to the conference, David Heinemeier Hansson gave his closing keynote for the day which was a great presentation about keeping the CRUD methodology across HTTP, defining proper models for has_many relationships using :through, and using Accept headers and file extensions to output the same data in different formats to different user agents. While talking about exporting objects to XML via APIs, he debuted ActiveResource, which is for the opposite end to take a resource like a 3rd party site's API and automatically turn the XML output back into a model that you can interact with at an object level as if the data came from your own database. Very cool stuff only two days into development according to David.