Walter notes two advantages of that scheme over extension methods. One of them is that extension methods can not be called with the standard prefix notation. This is not actually true, at least in the languages most known for extension methods, Visual Basic and C#.

That's not what I meant. PHP code was built around its SPL, if they accept nikic proposal, people will code using base types with similar operations set found in many dynamic languages.

What's left is [1,2,3]->slice(1,2) versus [1,2,3].slice(1,2)

ps: dynamic can be compiled and compiled can be interpreted, so I understand that it is unrelated.

The usual way to run Ruby on GAE is just to use JRuby on the App Engine Java runtime. Unholy has been dead for years, and from the commit messages seems to have never gotten to more than very basic functionality; I doubt very much that you could run meaningful code with it, on GAE or otherwise.

Having read Lisp in small pieces I can't hide that it was on my mind. But without parens smugness, right now php, python, es6 are getting closer and closer.