Interesting article on a high profile failure for Rails. There are two things that strike me here:
1) The idea of a single technology being a silver bullet. Thousands of projects have failed in the IT world when someone decided they wanted a clean slate with the latest technology (e.g Netscape). There’s mounds of data suggesting that refactoring is almost always a better solution.
2) Rails and Ruby have a reputation as slow. And it’s true. There’s performance gold to be mined in those hills. But so did Java when it came out. Java was slow as a dog but it was such a better language for writing applications that it quickly flourished and the speed improved dramatically. The Ruby world desperately needs a company which can put a team in place dedicated to speeding up the Ruby VM. There are efforts underway (e.g. Rubinius, YARV) but they are one man efforts (!) taking years to achieve. The one thing the open source world is bad at is solving problems which take man-years. But if the value is there, the problems will be solved.