Is Ruby Too Slow?

, Sep 14, 2006

Joel thinks so; David Heinemeier Hansson has a great pragmatic answer; Smalltalk guru Avi Bryant explains why dynamic, duck-typed languages don’t have to be slow at all, both he and Patrick Logan point out the link to Hotspot’s Smalltalk predecessor, StrongTalk; Steve Vinoski (who certainly has done his share of C++ development) criticizes Joel’s “fundamental misunderstanding”; Obie is disappointed by Joel.

I still think Joel is one of the greatest writers on the ‘net, but in this case, I agree he’s plainly wrong.

On September 14, 2006 5:41 PM, Jan Persson said:

So Joel is plainly wrong when he says that Ruby is slow? Have you seen http://shootout.alioth.debian.org/?

On September 15, 2006 12:31 AM, Stefan Tilkov said:

Ruby is obviously a lot slower than many other languages. That’s not the point. The major point is whether that maters.

The shootout doesn’t prove anything, IMO, since a Web app — the target of Rails — spends most of its time waiting on either the network or the DB. I’ve not yet seen a Ruby on Rails app that failed because of performance issues, and I don’t really expect to see one anytime soon.

On March 5, 2007 12:42 PM, raffaele said:

Please read this: http://joelonsoftware.com/items/2006/09/12.html

Joel is talking specifically about Ruby, not Rails. Ruby IS slow, fullstop. This may be secondary in webapps, but is still a matter of fact. This also means that Java is the best thing could happen to Ruby: rumors says that the JRuby implementation will be faster than the original c implementation.