ebay - cardamatic

From: Tony Duell <ard_at_p850ug1.demon.co.uk>
Date: Tue Feb 15 18:56:53 2005

> That's why you spent a lot of time desk-checking the program first
> ("playing computer"). And why if you found a bug, you tried to
> come up with a patch, rather than recompiling the program.
>
> Obviously now that compiling most programs is basically "free",
> desk-checking and patching have become lost arts.

The loss of desk-chcecking has, IMHO, led to a horrible method of
debugging which goes something like this

'Oh, it looks like an off-by-one error in that loop, let's change N to
(N-1) and try again'.

'Oh, and now let's replace that greater-than with a less-than'

'And ,,,,'

In the ned the program sort-of works, nobody really knows why, and the
result probably still has all sorts of bugs in it.

I've seen the same with hardware, particularly FPGAs (IMHO nobody should
be allowed near an FPGA until he cna design properly with TTL!). I've met
'desginers' who say things like

'Let's stick a D-type on that signal and see if it helps at all'

'I'll swap that NAD gate for an OR gate, it might do some good'

'Maybe inverting that clock will remove the glitch'.

Oh well....

-tony
Received on Tue Feb 15 2005 - 18:56:53 GMT