Subject: Re: Constructors motion 30 Version 2: small amendment
From: John Pryce<prycejd1@xxxxxxxxxxxxx>
Date: Tue, 7 Feb 2012 11:35:09 +0000
To: stds-1788<stds-1788@xxxxxxxxxxxxxxxxx>
P1788
Generally I have had a fair amount of agreement with the revised Motion 30.
I send it again with a small change suggested by Dan Zuras: the "virtual"
constructors have "bare" inserted in their names and the others have "dec"
removed from their names, since being decorated is the default.
Also shall num2interval(x) be abolished as Vincent suggests? Views please.
It can be re-invented at a language level.
John Pryce
I understand there is a certain danger to things like
num2interval(3.1). Still, known exact constants like
num2interval(2) are common enough& necessary. So I
would keep it. Perhaps the default decoration should
be different than nums2interval() but that is a question
for another day.
Which reminds me, either nums2interval()& its ilk
should take a decoration as a parameter or there
should be versions that do. And, of course, the
nature of that decoration is ambiguious at this time.
As usual, IMHO...
Dan
P.S. - While num2interval() takes a floating-point
number as input I'm sure that text2interval() can
be made safe by defining the output to be
[roundDown(text2num(text)),roundUp(text2num(text))].
(Perhaps you already have it defined that way.
I admit to not reading it closely. :-) But even
though I feel we need it, I cannot see a fool proof
way of protecting num2interval(). Alas, there are
always bigger fools.