Re: Stacking LSM proof-of-concept

From: Valdis.Kletnieks@private
Date: Sat Jan 29 2005 - 21:17:30 PST


On Sat, 29 Jan 2005 13:12:30 EST, John Richard Moser said:

> Uh.  I read my code like a thousand times, I was pretty sure I found all
> bugs.  I'm used to you know, not having any; but I'm used to writing my
> entire codebase from scratch, so I usually know what the *entire*
> program does.
> 
> I was surprised, is that wrong?

Yes, you were wrong.  The fact that one can produce a mostly bug-free codebase
from scratch (and I'll quickly point out the *vast* difference between
"bug-free" and "no reported bugs") only shows that you know how to produce
fairly small and self-contained programs.

Being able to write a 10K standalong program is *vastly* different than being
able to write 10K lines that will fit bug-free into a 10M line program.  And a
few days of lurking on lkml will show that *nobody* at the top is "used to not
having any" bugs - even Linus has at least one "OK, so I'm an idiot" posting
a month.

Even DJ Bernstein and Don Knuth don't claim to write totally bug-free code. ;)





This archive was generated by hypermail 2.1.3 : Sat Jan 29 2005 - 21:18:07 PST