Yesterday's blog post certainly generated a boat load of feedback, which I'm going to group into a few different tongue-in-cheek categories:
- Refactoring isn't a problem, just refactor everything yourself by hand, or write a refactoring tool yourself.
If I had the time, I would. Refactoring by hand is what I'm doing now on a daily basis, because I don't really have any other choice. Thus the request for better tools.
Which brings up a tangentially related point. Why is it that whenever someone asks for better tools it's because they're lazy? I think that we can all agree that it sucks to use an axe to chop down trees when you've used a chainsaw before. From now on, comments of the form "just use grep to do it" will only be accepted from those of you who program on mainframes and walk everywhere, simply because you remember the good old days before we had these fancy cars and trains to move us around.
- You must be coding Ruby incorrectly.
When making this comment, please point out the web page that shows how to write Ruby correctly. I have multiple Ruby webapps in production, I worked at Thoughtworks and I'm currently working with Pivotal Labs, two of the foremost Ruby consultancies, and no one has told me or my team that we're writing things incorrectly. However, I have an open mind and I'm willing to learn more about this topic.
- You can't just throw together Ruby code without tests, that's why you're running into problems.
Aren't you going to have this problem with any language? Where did I imply that this was something that we were doing? Anyway, 'rake stats' says:
Code to Test Ratio: 1:2.8
So I don't think that's the problem.
Seriously though, It's all a bit academic. If you asked me what framework I would use if I was going to start building a new webapp tomorrow (and told me that I couldn't use Seaside) I would still choose Rails. It's still the best tool for lots of jobs out there. Pointing out some of it's deficiencies doesn't make that any less true, despite what some people seem to think.