James, Roshan (Cognizant) wrote:

>//_I_ am not going to write an interpreter on .NET (nor on 
>//Parrot) but anyone else can.
>//
>//In the future, people may use Ruby on .NET or Ruby on Parrot 
>//daily, and my interpreter would be a reference implementation.
>//
>//							matz.
>
>Accepted Matz. It would also be unfair that everyone brings their
>problems to you and you are expected to keep solving them. I am looking
>forward to the ideas you put into Rite. 
>
>Maybe sometime Rite itself can be extended to bridge to .Net, such that
>Ruby is its best under Rite and the interop bridge is the only slow
>link. Sometime when you have designs for Rite available I would really
>like to see the features the runtime offers. 
>
>I however request you to put up reasons why Ruby wasn't targetted for
>the CLR sometime - even if that reason was simply 'I want to enjoy
>writing my own VM'. There is a reason why I am asking you for this.
>There are people at MS and ouside who are serious about getting the CLR
>(or CLI) to provide better support for functional and dynamic languages.
>Having a solid set of cases againt Ruby running on the CLR could direct
>sufficient research in that direction, which on the whole is a good
>thing. 
>
>Two things that bother me - 
>1) IronPython : Python on the CLR 
>	<http://www.hole.fi/jajvirta/weblog/20031210T0901.html>
>2) Groovy : A ruby like language on the JVM 
>	<http://groovy.codehaus.org>
>
>  
>
I have a student doing a proof-of-concept Ruby-to-CIL compiler. We 
expect to have something to show in late May. It is unclear if it will 
be of any use more than showing what is missing in CLR to support 
dynamic languages in a good way. OTOH, so far it looks rather promising; 
we even found a way to dynamically redefine methods without bending over 
backwards... ;)

Regards,

Robert Feldt