Re: Duplication of effort.

From: Harvey J. Stein <hjstein_at_bfr.co.il>
Date: 06 Oct 1998 23:30:23 +0200

"Perry E. Metzger" <perry_at_piermont.com> writes:

> Me:
> Perry:
> > > You'll never get a sufficiently rich standard.
> >
> > Why? It seems that C was. All that's needed is a standard FFI for
> > interfacing to C code. Then all the C libs/X/unix/everything else
> > come in for free.
>
> 1) The C standard is pretty easy by comparison. All that is needed is
> a definition of the ABI for function calls and you are pretty much
> set.

That's all that's needed in the scheme case.

In the C case the ABI for fcn calls is preset by the compiler & the
OS.

> 2) In the scheme case, if you can actually manage the spec, and to get
> it rigorous and interoperable, you have probably not only built a
> couple of systems to use it but have put more work into the spec
> itself than into any implementation of that spec.

The spec itself isn't difficult - the interoperability is
problematic. Look at Lars Thomas Hansen <lth_at_ccs.neu.edu>'s FFI stuff
& Bruno Haible's FFI lib.

-- 
Harvey J. Stein
BFM Financial Research
hjstein_at_bfr.co.il
Received on Tue Oct 06 1998 - 23:30:57 CEST

This archive was generated by hypermail 2.3.0 : Mon Jul 21 2014 - 19:38:59 CEST