[std-interval] Draft standard based on csets

[George Corliss]

All,

We have prepared a completely revised a draft of a proposed
standard based on the draft of March 2006 by Br\"onnimann,
Melquiond, and Pion. Although the attached document bears
the names of Br\"onnimann, Melquiond, and Pion, it is the
result of work by John Pryce on their earlier version, and
there has not yet been discussion on whether they accept
the document in this revised form.

The document is too large to send as an attachment to this
Alias.  The draft is available for download from
    http://homepage.ntlworld.com/j.d.pryce/isloct05/

This draft differs from the March 2006 draft primarily in its explanation of
an underlying abstract mathematical model based on containment sets and (we
think) simplifications which follow from that.  We have tried to strengthen
considerably the conceptual rationale for interval arithmetic.

This version has been the subject of extensive discussions among Interval
Subroutine Library (ISL) project members John Pryce, George Corliss, R.
Baker Kearfott, Ned Nedialkov, and Spencer Smith.  We intend this draft as a
catalyst for further discussion. We argue for the importance of specifying
an abstract mathematical model.  We favour a one-model cset-based standard,
but a standard that supports both this and the Simple System would be
acceptable.

In preparing this revision, the ISL project has focused in the interval
abstract mathematical model and the semantics.  We intend that C++ experts
in the community continue to refine the C++ aspects.

This draft does not reflect recent discussions of cmath or of comparison
operators.  We continue to follow those discussion threads with interest.

John Pryce, George Corliss, R. Baker Kearfott, Ned Nedialkov, and Spencer
Smith