Re: [Gnu-arch-users] Features command for arch

[Robert Anderson]

--- Original Message ---
From: David Allouche <address@hidden>
To: GNU Arch Users <address@hidden>
Subject: Re: [Gnu-arch-users] Features command for arch

>On Fri, 2004-09-03 at 07:36 -0700, Robert Anderson wrote:

>Please stop polluting this thread with your nonsense. You are
wasting
>everybody's time and your own.

My "nonsense" being:

1) putting the burden of reporting program variations on each
variant of the program is the wrong place for it. The reason is
that since each variation can't "know" how other variations might
behave differently, they can't sensibly know what to report about
their own behavior, as they don't know what other variations
might exist.  Only the consumer of a number of variants knows
what questions to ask, so the burden should be on the consumer,
not the individual variants.  The variants simply can't know, in
a fundamental way, what is relevant about themselves.

2) the 'configure' model for discovering a program's variations
in this vein (by the consumer, of course) has become ubiquitous.
 I think this is for good reason.

And to the retort: "But emacs has something like this (so it must
be right)"

Yes, but emacs is fundamentally different from tla in that emacs
interacts with a display in ways that are difficult or impossible
to discover by a shell script, whereas tla has no such
characteristics.

Just so people can decide for themselves if what I am saying is
nonsense.

As for who is wasting time in the sense of inhibiting tla's
progress by advocating going down the wrong path, I have a
different opinion than you.

Bob