Re: A suggestion regarding status change Motion P1788/0023.01:NoMidRad -- (VOTING PERIOD *mistrial!*)
Nate Hayes replied:
> Michel Hack wrote:
> > Nate must have overlooked the "Otherwise" part. Baker's
> > clarification was only about a possible mis-interpretation
> > of "non-standard", and did not change our collective position
> > on Kaucher arithmetic (discussed at length a year or two ago).
>
> I don't remember any votes on this.
> Apparently you have some "insider" information that I don't.
We never voted on this, that's true. What we did was to defer discussion
of support for any "non-standard" intervals, at a point (Nov 2008) when
different forms such as Kahan, Kaucher, Modal, Moore, Cset etc. were
indeed interfering with the most basic support that needed to be dealt
with first.
> > I also don't see how Kaucher intervals fit into Motion 19 -- I believe
> > the Kaucher issue was essentially settled by Motion 3, was it not?)
>
> Absolutely not.
Nate is right, and I was wrong there. The Kaucher interpretation of
intervals is indeed fully compatible with Motion 3; indeed, exclusion
of points at infinity was essential, as Nate reminded us back then.
Please forgive me for not remembering 150K lines of postings (3MB) --
and that is from a mail system that aggressively replaces quoted mail
with cross-reference pointers to the original, and eliminates long sigs
(except for first appearance) and gratuitous whitespace.
I am not familar enough with the practice of using Kaucher (or Modal)
intervals (or, frankly, even regular intervals, as (like Dan Zuras) I
came to this discussion more from a plain FP point of view) to judge
whether Kaucher arithmetic is ripe for standardization, and whether
such standardization should be pursued separately from P1788. What I
feel more strongly however is that, if P1788 were to standardize Kaucher
arithmetic, it should do so in a separate chapter or appendix, and to
present it as a separately-specifiable sub-standard (ooops -- bad choice
of words, as "substandard" has pretty negative connotations). My reason
for this is that the Kaucher interpretation is flatly incompatible with
other possible "non-standard" interpretations, e.g. Kahan-style, and it
should be possible to have a P1788-compliant implementation that extends
the basic intervals with those of a non-Kaucher flavour.
My reasons for wishing to avoid standardization of mid-rad forms is of a
different nature: this would weaken P1788, whereas properly-formulated
(and separated) Kaucher support might indeed strengthen the standard.
So, to get back to the (now in limbo) Motion 23: I strongly agree with
the exclusion of *explicit* support for mid-rad in THIS standard, but
see the exclusion of Kaucher arithmetic in a different light: it does
not belong in the main part of P1788 (which we should deal with first),
but could be considered later as an appendix, for example. (This is
also what happened to several pieces of 754-2008, for similar reasons.)
Michel.
---Sent: 2010-10-13 17:04:35 UTC