Re: Putting features in vs. leaving them out, especially functions
> Date: Wed, 25 May 2011 13:18:24 +0200
> Subject: Re: Putting features in vs. leaving them out, especially functions
> From: "Arnold Neumaier" <arnold.neumaier@xxxxxxxxxxxx>
> To: "John Pryce" <prycejd1@xxxxxxxxxxxxx>
> Cc: "stds-1788@xxxxxxxx" <stds-1788@xxxxxxxx>
>
> On Wed, May 25, 2011 10:06, John Pryce wrote:
>
> > - I rather disagree with George on reverse operations. Constraint
> > Programming seems
> > an important application (I know little about it) and it needs
> > operations of
> > the form "enclose the set of all x in xx for which f(x) is in yy"
> > where f is some
> > elementary operation. In CP applications I guess these are
> > performance-critical;
> > if so, that seems a reason to have them in the standard.
>
>
> > Stuff like (1-cos x)/x^2: people to whom these are
> > important are
> > competent enough to code them themselves; they are unlikely to be
> > performance-
> > critical.
>
> They are needed for the efficient implementations of slopes and
> second-order slopes, which we use in our COCONUT global optization
> environment almost as much as CP.
That is not really John's question Arnold. It can be
presumed that many things we do not specify will appear
in standard libraries that everyone uses. The question
is: Is it something fundamental enough to need to be
REQUIRED by 1788 or is it something that anyone can
correctly & efficiently implement given more fundamental
things that we DO require?
I expect that 1788 will have the equivalent of clause 9
in 754-2008. That is, a place where necessary but not
fundamental functions are listed so that their properties
may be specified in a standard way but none of the listed
functions are required for an implementation to be
considered conforming.
You (collectively) are more qualified to answer that in
the case of (1-cos x)/x^2 but that is the question before
you.
Dan