Just to nitpick, Objective-C does not have “named parameters”, it has selectors. [x a:1 a:-2 a:@"z"] uses a valid selector, and Objective-C and MacRuby both deal with that. Hell, even [x :1 :2 :3] uses a valid selector (no names for any of the parts), although I have no idea how either RubyCocoa or MacRuby solves declaring that. (It’s spectacularly bad form while overlapping with being valid.)

@Rafael – I’m actually in the midst of figuring that sort of thing out myself. It seems that the best way to go from obj-c => ruby is using the performRubySelector method, but unfortunately documentation is sparse right now (I believe the feature is brand new in MacRuby 0.4, so not too many people have played around with it yet.) However, I’m in the middle of a discussion on the MacRuby mailing list about it, so maybe I’ll get it figured out soon and create some documentation myself!

thanks for this great summary. I have a slightly different question: Is there any documentation / possibility to use Ruby classes and execute their methods right from Objective C? I’ve seen some examples that use NSClassFromString(), but they didn’t work in my case.

Very nice article! Thanks for taking the time to write this!

2 small things:

1) It is also possible to do NSTableview.new in RubyCocoa. However this example is confusing, since you are supposed to go through -[NSView initWithFrame:] when creating views.

2) If you want to overwrite the default -[NSObject init] method in MacRuby, you must still return self. However if you decide to overwrite #initialize (which -init calls], it is not necessary to return self here (as in pure Ruby).

Would you be interested to convert your blog post as an article for our website (http://www.macruby.org/documentation.html)? I think it would be awesome to have a “MacRuby for RubyCocoa programmers” short guide.

Laurent