Flavors (was: Motion 52: final "Expressions" text for vote)
On 2013-11-28 21:03:11 -0800, G. William (Bill) Walster wrote:
[About Section 7, on flavors]
> I thought Section 7 required the use of point function domains of definition
> based on the set of real numbers and/or its two-point compactification.
>
> Permitting a foundation based on the one-point compactification of the reals
> would be a step in the right direction.
>
> What about a foundation based on both the one- and two-point
> compactifications as well as other consistent systems.
I don't see why a compactification would be needed in P1788.
In general, compactification is not used in math, or it is
introduced for particular problems. This could be useful as a
flavor for some applications, but note that one- and two-point
compactifications are incompatible with each other.
> The point is that it is impossible to list the set of all possible flavors
> that might provide useful interval systems.
I don't see this as an issue. P1788 will define only two flavors.
If some other flavor is regarded as useful, then it could be added
as an extension to some implementation and/or be specified in a
future revision of P1788.
> So, why is it no better to impose the smallest possible set of
> constraints on flavors?
This may be difficult to do, with a risk of having too strong
constraints for future flavors.
BTW, should there be a function that returns information about the
standard and the current flavor, so that the user can test that the
implementation is using the flavor he expects? This would only be
a recommendation, as such information could be provided in other
ways by the implementation (e.g. in the flavor could be fixed in
some language, or it could be obtained from the function names,
or via macros, etc). I'm thinking of something like the __STDC__*
macros of the ISO C standard.
--
Vincent Lefèvre <vincent@xxxxxxxxxx> - Web: <http://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)