Re: Motion P1788.1/M005 to avoid non-conforming interval standards
On Sun, 7 Aug 2016 14:21:21 -0700 (PDT), Dima wrote:
> I believe that interval literal is the only topic where behavior is different.
The big deal is that 1788 allows <lower,upper> to specified in mixed radices,
and we want to avoid this in 1788.1 because determining whether lower<=upper
is difficult -- and in any case, 754 does not support mixed-radix comparisons.
We may have made a mistake in allowing mixed-radix interval literals in 1788.
In general mixed-radix support is easier in an interval standard than in a
pure numeric standard, because we can always enclose a result. It's even
easy to support tightest mode for basic arithmetic, except possibly for
the unlikely case where both decimal and binary mpfr-based types are used.
There should be no prohibition of mixed-radix arithmetic. But literals
are a different story, because text input is in fact unbounded-precision.
So we have an explicit escape clause, "unable to decide whether interval
exists", and I don't think we want that in 1788.1 -- it's complicated.
Michel.
---Sent: 2016-08-07 22:25:25 UTC