🚀 go-pugleaf

bevan@cs.man.ac.uk  writes:
> In article <39g66g$469@csnews.cs.Colorado.EDU> Tom Christiansen <tchrist@mox.perl.com> writes:
>    :-> In gnu.misc.discuss, bevan@cs.man.ac.uk (Stephen J Bevan) writes:
>    :In article <398vo8$ari@csnews.cs.Colorado.EDU> Tom Christiansen <tchrist@mox.perl.com> writes:
>    :   Translating ksh93, perl5, or python into scheme will only slow them
>    :   down.  
>    :
>    :Could you post the timings on which you presumably based this conclusion?
> 
>    As soon as you post the translators.
> 
> You made the claim, the onus is on you to back it up.  If you haven't
> got any timings, then you'll need to show that previous attempts at
> similar translations have consistently resulted in poorer performance
> -and- that this is inherent in any such translation.  I don't believe
> the evidence is available to show the former let alone the latter,
> but I'm perfectly willing to be proved wrong.

As to whether translating from one language to another will slow things
down, let me note that it depends how it is done. For example, I tend
to use a FORmula TRANslator to do scientific programming in Forth (written
in Forth, natch, so it is only 7 K or so including a 1K buffer). The
code that employs such would look something like

	: example  F"  a = (b+c)*EXP(d+e)/ACOSH(w) "  ;

The nice feature of this is it looks to me--an old FORTRAN curmudgeon--
like what it is, rather than like a reverse-Polish gumbo. So the source 
is readable, granted one understands the usual Forth conventions.

However, if I decompile the above example, I will see only the
Forth that the formula was translated into--in fact, an RPN gumbo.
That is, nothing whatever was lost in the translation--it will execute
as fast as though I had not imposed a FORTRAN-like interface. 

Although not a Scheme user, I believe it would be just as feasible to 
do this in Scheme, and that no runtime penalty would be imposed, unless
the translator produced worse than average Scheme translations of its
input.


--
Julian V. Noble
jvn@virginia.edu