Re: F22 Raptor in slashdot.org

[aek@vib.usr.pu.ru (Alexander Kopilovitch)]

tmoran@acm.org wrote:

>   Programmers tend to be smart and try very hard to be logical and still
> have an error rate of x/SLOC.  Mathematicians try *really* hard to create
> logically correct proofs, yet a published result is still sometimes found
> to be incorrect.  Is there any reason to believe that decision makers,
> managers, military officers, politicians or any other group has a
> substantially lower error rate?

Generally yes, one may have such a reason: members of the latter group quite
often are backed with a staff of aids and/or advisors of various kind.

Then, the notion of an error for all 3 metioned groups is different. It may
surprise you, but even for mathematicians this is not always simple: for example
consider Euler's claim that the sum of all natural numbers is equal to (-1/12).
Was that an error? What seems a nonsense in 18th century and clearly an error
in 19th century, appeared a brilliant prophecy in 20th century. Well, it is true
that such a case is very rare in mathematics; but for decision makers, magagers,
politicians and even for military officers this isn't so rare: explicit decision
may seem obviously wrong, but its consequences appear surprisingly good - so,
was there really an error?

Finally, what really matters here is not errors themselves, but excuses
for the members of the latter group. Excuses not only for the front-line
error-makers, and even not for their staff members, but most surely for their
"human programmers" - those who educated them, drilled them and consulted them.



Alexander Kopilovitch                      aek@vib.usr.pu.ru
Saint-Petersburg
Russia