[gclist] Concerns about GC

[Henry G. Baker]

> 	The problem is that GC does not make economical use of a
> valuable resource: memory. The initial start-up requirements for most
> GC implementations is 8M of memory, though more often 16M. Everytime a
> program requires more than 16M, the system gobbles more memory and
> never gives it up (I think Emacs returns free buffer space to the OS,
> though).

I think that you have fallen prey to a logical fallacy:

A.  Large, complex programs use of lots of memory.
B.  Large, complex programs find GC useful & helpful to manage lots of
    memory.
C.  Therefore, the use of GC => the use of lots of memory.

Now, this is an almost valid inference using Bayes theorem, since most
small, non-complex programs won't bother with GC, but there is no
necessary logical relationship between 'large/complex' and 'uses GC'.

-----

I could make roughly the same argument for 'expensive, sophisticated car'
and 'uses an automatic transmission'.  Unfortunately, the existence of
expensive sports cars with manual transmissions screw up this inference.

-- 
Henry Baker
www/ftp directory:
ftp.netcom.com:/pub/hb/hbaker/home.html