RE: bytecode unification for scripting languages

["Will Fitzgerald" <fitzgerald@inetmi.com>]

Just as a reminder: There's a .NET Scheme project (based on Scheme 48) at
Nortwestern by Pinku Surana under Ian Horswill's direction.

See: http://www.cs.nwu.edu/~surana/


> -----Original Message-----
> From: owner-plt-scheme@fast.cs.utah.edu
> [mailto:owner-plt-scheme@fast.cs.utah.edu]On Behalf Of Shriram
> Krishnamurthi
> Sent: Tuesday, July 31, 2001 9:03 PM
> To: plt-scheme@fast.cs.utah.edu
> Subject: bytecode unification for scripting languages
>
>
> I would be surprised to see continuations in such a common runtime.
> And even if they were there, I suspect there would be many tricky
> issues involved in supporting them consistently across languages.  I
> doubt "stackless" Python has DYNAMIC-WIND, which fundamentally affects
> the implementation of CALL/CC.
>
> In the end, what's the point here?  Microsoft's .NET is already
> shipping, and they will be releasing a sorta-kinda-open source
> implementation next year.  Sure, it's not going to be as fast as the
> native Windows version, but I have no reason to believe yet another
> soaker-upper of hacking resources (Parrot) is going to attract the
> world's great optimizer writers.  More likely the good optimization
> work will happen for the CIL.  So this will merely create a new,
> incompatible run-time system.
>
> Why not take the CIL/CLR spec and build a really great implementation
> of it instead?  Here's an opportunity where Linux and the Open Source
> movement can completely embarass Microsoft.  Be the first to provide a
> free and openly hackable implementation of it, well before Microsft
> does, and make the Linux version run five times as fast as the Windows
> version.  Now *there* would be something impressive.
>
> Doubt it'll happen.
>
> Shriram
>
> PS: Full disclosure: I just spent 3 days at MS hearing about .NET.
>     But my comments are based not on the fact that I'm easily bought
>     -- perhaps I am -- but rather than the CIL/CLR is, at least on
>     paper, *good*.  The Open Source movement has never led the way in
>     the design of *anything*, anyway, afaik.  Why would I assume
>     Parrot would be better in any sense other than the strictly
>     ideological?
>