Re: Re: Meta-level questions

Mark Hahn wrote:
>...
> 
> I didn't mean to imply that this was open for voting.  I have learned from
> this whole experience that my number one goal is for Prothon to have a
> community and having fun with new features is secondary.  

If that's your goal then I think that the the .NET option is better. You 
will immediatley have access to the community of .NET programmers with 
few dynamic language options. If Prothon uses C as its runtime then you 
are sharing that community with Ruby, Lua, Python, REXX, ... It seems 
like a much more crowded marketplace.

 > ...
> If you think about it, it is not too surprising that Gudio would think I am
> doing the wrong thing.  I am deviating from what he did.

He didn't really have a choice. Python predates both the JVM and the 
.NET runtime. But yes, he has pretty much decided that Python 3.x will 
have its own VM and of course he would love to have Prothon test out 
some VM implementation ideas.

I see a risk going purely with .NET/Mono or purely with your own VM. You 
don't have control over .NET/Mono and Microsoft may take it in 
directions you don't like (e.g. there is a small chance that they sue 
Mono out of existence). On the other hand, it might be the case that in 
five  years the very idea of a language that runs on its own VM might 
seem anachronistic, like a language that comes with its own operating 
system (remember those?) or CPU (ditto).

Perhaps the best policy is the following:

  * all major scripting languages in the future will have a .NET 
implementation, JVM implementation and close-to-the-metal C 
implementation (if there are performance benefits to doing so).

  * It is easier for Prothon's first implementation to be built on a 
runtime that already does multi-CPU JIT[1] and a standard library[2].

  * After that first implementation is done, then we can do a 
knowledgable evaluation of what the benefits of a pure-Prothon VM would 
be. We'll be able to say with confidence: "we know this is worth doing 
because we've pushed .NET to its limit and we can do better." If Prothon 
could benefit from going beyond the limits of .NET then great: we'll 
know how and why.

  Paul Prescod

[1] http://www.go-mono.com/ppc.html

[2] http://www.onlamp.com/pub/a/onlamp/2004/03/11/mono.html