I’m starting to see more startups using Ruby and Ruby-on-Rails for three typical types of projects:
- Simple web sites, e.g., the 37signals properties.
- Web 2.0ish sites where XML will be passed back’n’forth for both AJAX and site integration purposes.
- Internal use around flexible scripting, product extensions, release engineering, etc. For example, one of my startups still in stealth is going this way.
Uses (1) and (3) are right down Ruby’s and Ruby-on-Rail’s alley. My personal experience is that Ruby and Rails are very productive–I really enjoy them.
I thought Tabblo was headed towards (2) but it turns out they decided to not use Ruby on Rails (see comment from founder Antonio Rodriguez).
I’d be a little concerned about the risk of doing (2) without an absolutely top-notch team that has connections into the Ruby community. The risk is three-fold. First, the framework APIs are fluid and the documentation is poor. Second, as noted by Tim Bray, Ruby’s XML/Web services support leaves a lot to be desired. Third, the collection of Ruby libraries wrapping 3rd party systems & services is still immature. That’s why I see Ruby as a great choice primarily for self-contained apps right now.
Hey Sim,
We actually don’t use RoR at Tabblo. Last summer I spent some time looking at, and even like it enough to run my blog on it, but I wouldn’t build a site that is expected to serve millions on users for the following reasons:
1. Ruby is weak-ish on XML, internationalization, and general optimization hooks for the runtime.
2. Rails is a little too tightly integrated for my tastes.
3. It’s also got a bit too much magic for me but I realize that this is what people tend to like about it.