[plt-scheme] to define, or to let

From: Bradd W. Szonye (bradd+plt at szonye.com)
Date: Sun Mar 21 17:27:20 EST 2004

Anton van Straaten wrote:
>>> I'd encourage taking standards seriously, too.

Matthias Felleisen wrote:
>> This is only true for standards that are up to certain standards.

I agree, and I believe that R5RS is sufficiently good to follow.

>> In case you didn't understand my previous response, I don't think
>> that this is the case with R5RS. And just as with run-away regimes,
>> such times call for civil disobedience.

R5RS is about as good as the C standard, which is plenty good enough for
engineering use. Your standards may just be higher than most folks'. And
don't forget that language standards cater mostly to the needs of
typical engineers and programmers. If you want something rock-solid for
academic or life-critical use, you need a better standard.

> I have a different perspective on R5RS.  It's unusual for a language
> standard, because there's no question it defines a family of languages
> which have some significant differences among them.  It's not perfect,
> and it's not complete.  However, some of the incompleteness has to do
> with not overspecifying and unnecessarily constraining
> implementations.

In short, it's a good language standard for engineers. Folks with higher
standards, like Matthias, would probably prefer a good language standard
for academics, which is something entirely different.

> Some have claimed that this is a post hoc rationalization of a
> standard which evolved as an uneasy compromise amongst differing
> implementations.  I maintain there's value in the result,
> nevertheless.

That's how most language standards work, IME.
Bradd W. Szonye

Posted on the users mailing list.