//stack based languages do computation on stack. They usually
//identify var as index and process the stack. variable names
//vanish. Ruby, I think will need the variable names at many
//places to process code.
Aah, yes, variable names are lost for local variables - but member
variable sof classes do retain their field names. So for most dynamic
behavior that would suffice - when there are behaviors that involve
something like an “eval” changing a local variable I would expect the
compiler to generate some sort of plumbing code that will make
appropriate modifications to the local variable on the stack.
//However, for a majority of Ruby’s dynamism we have spikes
//that show how
//to solve them; for the rest I’m more skeptical. In
//particular, I don’t
//see continuations happening, threads will be different and
//probably some
//other things I forget about now… 
//David Simmons, the creator of Smallscript/S# [0] (Smalltalk
//for .Net),
//used to frequent this list. I believe he has discussed many of the
//issues with porting a dynamically-typed language to the CLR (some on
//ruby-talk, but also at [1]).
//
//[0] http://www.smallscript.org/
//[1] http://www.smallscript.org/Downloads/SSharp_NET_Notes.asp
Thank you for being open enough to discuss this technically - I was
afraid of this thread degrading into mindless flame wars.
Robert, Sir, can you tell how I could know when your project is released
- do you have a notification system or something similar? I would be
very interested in your work. I cannot seem to find a project page about
this, is it this one?
http://www.ce.chalmers.se/~feldt/ruby/ideas/rubyvm/
Regards
Roshan
InterScan_Disclaimer.txt (520 Bytes)