Project 4: Minimal hardware requirements

A product of this project "Minimal hardware requirements" was the
article at the end of this file. No official position on it was
agreed by the group.

---------------------------------------------------------------

Oxford 1975:
============

On computer hardware it was agreed that Dr Reinsch should act 
as co-ordinator.

Oak Brook 1976:
===============

Dr. Reinsch reported on progress of his project on Minimal 
Hardware Requirements. He felt that, although related to machine 
characteristics, it was a far more difficult topic because it was 
easier to describe existing hardware than to design ideal hardware. 
He would be reporting in detail on the topic during the Portability 
Workshop and it was agreed to hold a session on minimal hardware 
requirements on the Wednesday afternoon following the discussion of 
machine characteristics. Dr. Reinsch was asked particularly to 
discover the views of Professor Kahan. The Group would return to 
the topic at a later session.

Amsterdam 1977:
===============

Since Dr. Reinsch was unable to attend the meeting, Professor Hull had 
agreed to present a report on work associated with the development of 
specifications for desirable floating point facilities.

Jeremy Du Croz's report on the Oak Brook meeting that was chaired by 
Dr. Reinsch was available to the Working Group, and the 
first part of Professor Hull's presentation was to summarise briefly the 
contents of this report. He then gave the following reasons why he believed 
it would be worthwhile continuing with this work:

(a)	there seemed to be a considerable degree of 
	agreement on the topics considered so far;

(b)	the need for standards is great and the potential 
	payoff, if a responsible group such as WG 2.5 and 
	other interested persons should reach agreement
	on there topics, would be very great;

(c)	hardware changes are taking place (e.g., hand 
	calculators, MU-6) and there does seem to be a 
	chance of influencing such changes.

He then went on to outline steps that might be taken which would ultimately 
lead to an appropriate document.

The Working Group agreed that discussion of the topics should continue 
informally between members of the Group outside the regular meeting times of 
the Group (and subsequently an evening was devoted to this subject). It was 
also agreed that Professor Hull would prepare a document based on the report 
of the Oak Brook meeting, and the discussions just referred to, as well as 
on his own subsequent work on the subject. This document would be circulated 
to the Working Group, and also to a few other selected individuals 
(such as Dr. W.J. Cody and Professor W. Kahan) well before the time of the 
next meeting. It was hoped that a revised version of this document could 
be published fairly soon, in, for example, the SIGNUM Newsletter, in order 
to give the ideas wider circulation and to get feedback from other interested 
persons. If all went well, the Working Group could look forward to the 
possibility of producing some kind of "official" document.

---

A document would be prepared by Professor Hull. This would be sent to 
members of the Group, attendees at the meeting at Oak Brook, Professor Edwards 
and Professor Wheeler for comment. If there was an agreed consensus the 
document would then be given a wider distribution. Failing that it would be 
discussed in detail at the next WG 2.5 meeting.

Madison 1982:
=============

It was noted that the first hardware features project has been
superceded by the second.

Document:
=========

C.H. Reinsch: 'Principals and preferences for computer arithmetic',
SIGNUM Newsletter 14 (1979), 12-27.