Costs of optimization (was: Re: Information security, recycling and irony)

bmcculley at rcn.com bmcculley at rcn.com
Fri Feb 3 23:55:01 EST 2006


Fred wrote:

>Just a little thing like using *real* recycled paper, not 
>reused paper, with perhaps their logo printed on it as an 
>added touch, would not cost all that much more AND would 
>avoid snafus such as this.

"would not cost all that much more" is the issue.  Our free
market economy has gotten so efficient that *any* additional
cost is fatal, and *any* cost-cutting is vital.  So the
benefit of re-using waste paper as wrappers is seen as compelling.

UNTIL...

the cost of compromising confidential data kicks in.  That's
an unintended and unforeseen consequence though, so the bean
counters haven't figured it into their models.  What is
needed, and not yet considered, is some way to include the
costs when optimizations break.  Usually it seems what happens
is that optimized performance reduces costs fractionally, but
failures incur costs increased by orders of magnitude over the
costs of operating a simpler but more robust system.  But
that's just an intuitive appraisal, has anyone done any formal
research into these issues?

-Brucem



More information about the gnhlug-discuss mailing list