ghc image size

Stephan Reiff-Marganiec [email protected]
Wed, 19 Dec 2001 09:34:18 +0000

At 10:51 19/12/2001 +0200, you wrote:

> >> Why is executable size a barrier?  1.64 megabytes (that's the size of
> >> the executable I built with GHC most recently) of disk space costs less
> >> than half a cent.
>  AL> I don't like this argument.  Can I go to a computer store, pay a cent,
>  AL> and get a hard disk with space 1.64 megabytes or more?  Until then, I
>  AL> can't believe that 1.64 megabytes of disk space costs less than half a
>  AL> cent.
>1 year ago I bought 20Gb for 200$. That will be 200/(20*1024) = 0.009765625
>dollars per megabyte. Pretty much less than half a cent.

I don't think it is about how much memory or hard disk space cost. If it is 
possible to reduce the footprint of any software without reducing 
efficiency than this is certainly worthwhile.

I believe that there is a more scientific issue behind this, as was 
correctly pointed out by Albert. It should be investigated why the 
executable is so large, and then considering whether it can be done in a 
better way. Just throwing more hardware at things does not mean that they 
are better and certainly seems a very short sighted option. Maybe there is 
a genuine problem there and the executable cannot be any smaller, in which 
case this is good to know (and might initiate some nice research). If it 
can be done smaller, what is wrong with doing it?


Stephan Reiff-Marganiec
Research Fellow
Department of Computing Science; University of Stirling
email: [email protected]    tel: 01786 46 7448