Peter Bex scripsit:
> What I also don't understand is why we keep inventing (mostly) brand new
> things. Earlier someone mentioned that the Scheme *report* is a report
> on existing practices, in order to standardize those things that have
> been proven to work in practice. I think this is a good goal to keep.
I agree entirely, in principle. However, sometimes standards committees
do have to invent things, a practice with a mixed record. The X3J11
committee that standardized ANSI C (now ISO C '89) invented type
signatures for functions, and all cheered. C programming is difficult
enough, and without type signatures it was that much worse, so they were
quickly adopted.
They also invented the infamous "noalias" qualifier: Dennis Ritchie
thundered from Olympus[1], and "noalias" was quickly retracted, never
to be seen again. A third invention, trigraphs, whereby people could
write ??< and ??> if they didn't happen to have { and } on the keyboard,
went into the standard, turned out to be un-useful, and sank like a rock;
even gcc only implements them if you say -std or -trigraphs.
[1] http://www.lysator.liu.se/c/dmr-on-noalias.html
> Let implementations experiment with this idea and when 2 or more
> implementations have come up with something that really works and
> when all the details have been ironed out and people understand
> the consequences, standardize it. This also helps us keep the WG
> focused.
That is exactly what I have asked the SC to do: not let any features
in without at least two implementations. But that does not mean
that the two implementations had to have the feature *before* the
process begins! Standardization will take years (hopefully a small
number of years, but still years).
--
John Cowan [email protected] http://www.ccil.org/~cowan
Thor Heyerdahl recounts his attempt to prove Rudyard Kipling's theory
that the mongoose first came to India on a raft from Polynesia.
--blurb for Rikki-Kon-Tiki-Tavi
_______________________________________________
r6rs-discuss mailing list
[email protected]
http://lists.r6rs.org/cgi-bin/mailman/listinfo/r6rs-discuss
