[comp.compilers] translations

kym@bingvaxu.cc.binghamton.edu (R. Kym Horsell) (09/19/90)

In article <1990Sep18.103653.24032@dce.ie> ch@dce.ie (Charles Bryant) writes:
>That's funny, I can't see any bug in the original. I can see a bug in the
>Pascal one though: when the list is empty new elements can't be inserted
>because 'head' is never changed.

The ``bug'' was the lack of initialization (_not_ really a bug
in C but it _is_ so in other languages, and therefore in general).

It was a rotten trick anyway. If you initialize ``head'' to 0
and allocate ``mem'' cells starting at 0 the Pascal program's ok.

As for saying ``well ok -- _try_ to perform the translation using
Pascal pointers''. This is the same as asking ``translate this
program, that uses language feature F, into one that _doesn't have_
feature F''. A non sequitur?

As I said in the original post, matters of _efficiency_ are quite distinct 
from the original statement that "X->Y is easy and Y->X is a hard 
translation".  

>From the point of view of formal language thy _all_ translation is _very hard_ 
(i.e.  ``impossible'' and therefore equally ``easy'').

Saying that a Pascal program that implements lists using integers and 
arrays is either _time_ inefficient and/or _space_ inefficient is beside 
the (original) point.

In any case, efficiency is (usually) a matter of a particular underlying 
architecture (WELL off the track of the original question). On some systems 
using an array is actually preferred to manipulating pointers; the former is 
under more-or-less control, whereas the latter is subject to the vagueries of 
the C memory management. It may be possible to maintain locality-of-reference 
with integers and arrays and ``cache misses'' can kill dominate program
performance considerations.

Also, on VM systems allocating ``huge arrays'' does not present such an 
overhead as you seem to think. In fact, the C malloc routine on _some_ systems
essentially declares a _huge_ static array to which it passes out pointers;
the VM mechanism does the rest.

-Kym Horsell
-- 
Send compilers articles to compilers@esegue.segue.boston.ma.us
{ima | spdcc | world}!esegue.  Meta-mail to compilers-request@esegue.