His main argument is that performance will suffer when you run a program using an interpreter which itself is running on a bytecode virtual machine (his exact words), which is true.
So I won't do that. 
There is no interpreter involved in running ActionScript on Flash 9 player today, and there won't be one in IronMonkey.
Note: you don't need to cite YouTube or EVE Online to justify Python (even if they are real, they may not seem real to non-believers . It's enough to point out that CPython easily outperforms SpiderMonkey, the current JavaScript engine used by Firefox.
AFAIK Tamarin gets its speedup thanks to optional static type declarations. So I don't know how it would help the ironpython port...
(Paraphrase) AFAIK C# running on CLI gets its speedup thanks to its static type declarations. So I don't know how it would help IronPython running on CLI... Oh, I do know.
Hum. The main problem with performance here is how good the concepts from CLI maps to concepts of tamarin. If they map good, than it should be unnoticable (provided that tamarin is indeed fine-tuned, which I've got no idea about). My concern is that those might not map very well, which will mean performance penalty with approach of translating CLI bytecode directly to tamaring bytecode. But well, time will tell.
PS. Why comment window is so terribly small?
The Mono project gets a double dose of hate from some vocal open-source advocates. Even more hate than Microsoft gets.
> But well, time will tell.
Every time you hedge your bets, a baby angel loses its wings. Please be more careful in future. Think about the angel babies.
Commenting by HaloScan
