Re: [boost] ML Interpreter for Template Metaprogramming

From: Vesa Karvonen (vesa_karvonen_at_[hidden])
Date: 2002-10-03 14:06:47

David Abrahams:
>I think the only really "un-cool" thing about your language is the number
>of tweaky abbrevs. It for me it cuts the readability a lot to truncate
>5-letter words by two letters.

I like short names..., but I have renamed the constructs to use 
non-abbreviated names.

Larry Evans:
>I'd be interested.  I've had trouble understanding some of my template 
>code.
>I'm hoping that what you've done my help.

I have uploaded the prototype ML interpreter to the Files section. You can 
find it at here:

  http://groups.yahoo.com/group/boost/files/ML_Interpreter/

The test should compile (it'll take a long time) directly on g++ version 
2.95.4 20011002 (Debian prerelease) - with a sufficiently high template 
recursion depth. I have not tested the code on other compilers and it is 
possible that it is not 100% ANSI.

My main motivation was to find a convenient way to express algorithms. 
Interpretation makes this technique very slow. Some speed-up could be 
obtained by further optimizations, but probably not enough.

David Abrahams:
>This reminds me strongly of Spirit subrules. Were you inspired by Joel's
>work on Spirit?

No, the language constructs were borrowed mostly from Lisp and Standard ML / 
Ocaml. The idea of building an interpreter to allow convenient syntax popped 
to my mind when I read a couple of messages on the Boost list regarding the 
strict evaluation problem. I have recently worked on a couple of small 
compilers and interpreters for functional style languages.

_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com