[Cialug] Nuremberg Defense for Standards

[Todd Walton]

On Tue, Feb 17, 2009 at 7:00 PM, kristau <kristau at gmail.com> wrote:
> Most likely the ISP and Windows are wrong compared to the [---]
> standard and Linux is doing it right.  But because both the ISP and
> the Windows box are ignoring the standard, it is Linux that "appears"
> broken.

Speaking generally, it seems like sometimes "the standard" is used as
a Nuremberg defense for poor performance.  I support the use of
standards.  Advantages include having your process well documented
(your documentation is written for you!), most other systems will
already interoperate with you (theoretically) and, besides, making
un-standard improvements to your system is sometimes just a mask for
the real problem instead of addressing the problem outright.
Standards provide a meeting point for constructive debate.

But if Linux is going to follow a standard, then it must also take
responsibility for the standard.  If the standard is broken, then
standard-following Linux is broken as well.  We don't get a pass by
"we're just following the standard".

This is just speaking generally.  This may or may not be relevant to
the specific case being discussed, DHCP.  Also, a related but distinct
discussion would be "standards that are perfect if only the world were
perfect" or "standards written for an imaginary universe".

-todd