[LMB] OT: Comment markers (was Best One-Liners)

Michael R N Dolbear m.dolbear at lineone.net
Sun, 17 Feb 2002 15:06:17 -0000


> From: Damien R. Sullivan <phoenix at ugcs.caltech.edu>
> Subject: Re: [LMB] OT: Comment markers (was Best One-Liners)
> Date: 17 February 2002 05:28
> 
[...] 
> I'm no MS fan, but I'll speak up here.  "63,000 known problems" doesn't
> tell as much.  What's the threshold?  Official bug reports?  Any flaw
> they're tracking?  Software only, or software and documentation?  If the
[...]
> On the flip side, how does one count a basic design flaw such as
> "doesn't use basic memory protection and applications can crash the
> computer"?

My take on this is that defects in the initial specification are not bugs
unless they can be removed without causing an incompatablity. Some systems
were famous for their "designed-in" bugs.

Anyone involved with mainframe CICS will remember that one application
could stop all others from making progress. Getting round this was one
unexpected advantage of multiple-cpu machines.

OTOH perhaps a crash/freeze rate due to sofware has an optimum value of
about one/week, otherwise humans get into bad habits and/or forget how to
deal with hardware or power supply caused failures.

Little Egret