[linux-elitists] Avoiding Mailman Duplicates

Rick Moen rick@linuxmafia.com
Sat Feb 21 09:31:33 PST 2004

Quoting Lance Simmons (lance@lsimmons.net):

> Standards emerge, don't they, as more and more people see the advantage
> of doing things a certain way.  Those advantages include the fact that
> others are also doing things that way.  [...]

Standards documents clarify and codify practices that are believed to be
feasible and reasonable.  Generally speaking, successful standards
documents tend to abstract from, codify, and at most only slightly
extend existing implementations in actual use.  SMTP happens to be an
example of that.

The "standards from on high" crowd apparently we _will_ always have with
us:  That disease affected even the POSIX standards effort, some Linux
old-timers will recall.  We can be thankful that Torvalds did _not_
attempt to implement all of POSIX.1, but rather let it be known that
he'd be happy to include support code for any POSIX calls actually
needed in real implementations.  And he did exactly that -- and no
further.  Our need for gratitude owes to the resulting lack of pointless
cruft in the kernel to implement the extremely numerous parts of POSIX.1
that _nobody ever uses_.

And here's the result:  When the Austin Committee recently started
revising POSIX, they invited a number of Linux kernel coders, primarily
in order to leverage their experience to _remove_ pointless crud that
nobody had ever used, that had been lodged in the spec solely because
some feature-itis-obsessed committee member put it there.

> Does that also strike you as quaint....

No, just incomplete.

> ...or did the poster at whom you are sneering saying something
> stronger than this?

Rather the latter, I would say.  Someone who says (by implication) "I'm
not interested in a technical practice until it's in an IETF-adopted RFC" 
has shown himself conclusively Unclear on the Concept.

Rick Moen                     Emacs is a decent operating system,
rick@linuxmafia.com           but it still lacks a good text editor.

More information about the linux-elitists mailing list