Top 500 trends

Josip Loncaric josip at icase.edu
Wed Nov 27 08:45:12 PST 2002


John Brookes wrote:
> 
> [...] don't be so down on building now [...]

Agreed.  We cannot expect the future to automatically make everything better. 
We must actively push and mold this future, or else it won't happen.  

Instead of laboriously computing Pi by hand in ancient times, should humanity
simply have waited a few thousand years for a cheap electronic calculator with
a "Pi" button"?  It is a Good Thing that the ancients had NOT waited, because
without their efforts we would not even have modern computing.

When we seek to minimize the cost of a computation, the temptation is to
simply extrapolate Moore's law and postpone the effort as long as possible. 
However, the value of the computed results depends on time.  A decent
approximation to the correct answer computed today is usually much more
valuable than a high fidelity result obtained at the same cost a decade later.

This fact is essential for progress.  Without the pressure to build NOW, where
would future computing advances come from?  If everyone stopped buying
computers in expectation of faster ones to come, Moore's law would end.  In
this case, patience is not a virtue...

Sincerely,
Josip


-- 
Dr. Josip Loncaric, Research Fellow               mailto:josip at icase.edu
ICASE, Mail Stop 132C           PGP key at http://www.icase.edu./~josip/
NASA Langley Research Center             mailto:j.loncaric at larc.nasa.gov
Hampton, VA 23681-2199, USA    Tel. +1 757 864-2192  Fax +1 757 864-6134



More information about the Beowulf mailing list