[std-interval] Comments on the Interval Standard Proposal

[Sylvain Pion]

Gabriel Dos Reis a écrit :
> Sylvain Pion <Sylvain.Pion at sophia.inria.fr> writes:
> 
> | Gabriel Dos Reis a écrit :
> | > | For implementations which do not support infinity, how do you check if
> | > | an interval is bounded ?  Currently we can't.
> | > Hmm, I'm unclear about this one.  If the implementation does not
> | > support infinity, does not it follow all intervals are bounded?
> | 
> | I was thinking that implementations could store extra data members
> | specifying if a bound is infinite.  You can't return it through
> | .lower(), because the floating-point type does not have a value
> | for infinity, but is_positively_bounded() can return the
> | information.
> 
> Aha.  So, is it accurate to summarize the issue as to how to support
> unbounded intervals even where the implementation has bounded
> arithmetic for scalar arithmetic types?

Basically, yes.  So far the proposal aims at putting our head deep
in the sand and forgetting the issue, praying that users of such
implementations won't really use intervals.  But now that Guillaume
wrote that we were almost there, I thought we could push it a bit
more and provide a correct standard (respecting the basic inclusion
property even for overflow cases).  If it comes easily, then let's
do it.

I just hope that nobody will ask for integral types now... :)

-- 
Sylvain