Milestone-Proposal talk:IEEE Standard 754 for Floating Point Arithmetic

Original Citation Title and Text -- Administrator4 (talk) 19:26, 19 September 2022 (UTC)

IEEE Standard 754 for Binary Floating-Point Arithmetic, 1985

In 1978, faculty and students at U.C. Berkeley drafted what became IEEE Standard 754 for Binary Floating-Point Arithmetic. Inspired by ongoing collaboration with Intel, the proposal revolutionized numerical computing. Its robust arithmetic and standard data types promoted unprecedented program reliability and portability. By 1980, microprocessor companies were already implementing the proposal, which was formally approved in 1985. IEEE 754 remains the standard for reliable numerical computing.

Re: Original Citation Title and Text -- Bberg (talk) 22:15, 19 September 2022 (UTC)

As Advocate, I accept this citation, and look forward to this advancing to consideration by the History Committee in October.

Brian Berg

Re: Original Citation Title and Text -- Jeromecoonen (talk) 17:11, 20 September 2022 (UTC)

On behalf of Co-Proposer James Thomas and myself, we accept this citation and gratefully encourage the History Committee to consider it for approval in October.

Jerome Coonen

Advocate Decision re: this Milestone Proposal -- Bberg (talk) 19:00, 19 September 2022 (UTC)

This proposal is an excellent example of clarity and a depth of documentation of background information. I note that every word of the citation is explicitly discussed and supported, and that the subsequent information in the proposal nicely fleshes this out. I thus fully support this proposal.

Please note that the 2 Expert Reviews have been included on the proposal page as opposed to this page, and that this will allow the insightful comments included therein to appear in the final Milestone page.

Brian Berg, Advocate for this Milestone proposal

Re: Advocate Decision re: this Milestone Proposal -- Amy Bix (talk) 17:39, 23 September 2022 (UTC)

I think the citation is indeed well written. My only question would be with the last sentence, "IEEE 754 remains the standard for reliable numerical computing" - can we guarantee that this will still be true for someone looking at this plaque in another 30 years, 50 years, or more? If not, maybe rephrase with something like "IEEE 754 remained the standard for reliable numerical computing over following decades."

Re: Re: Advocate Decision re: this Milestone Proposal -- Jeromecoonen (talk) 20:14, 23 September 2022 (UTC)

Thank you for your comment, Amy Bix. We've had much discussion of the existential element of a plate of bronze cast in time, and yours is a legitimate concern.

For truly past achievements, a plaque can represent the facts best it can, for future generations to ponder.

For an ongoing and vital milestone, like the floating point standard, we cannot predict the future, but I think we seriously weaken the message when we introduce language like, "remains the standard up to this casting". (I'm overstating your case.) Does not the date on the bronze set the context adequately, so that we can make the cleaner statement?

The reality is that 754 does remain the standard, whose influence is now felt in expanding areas like Machine Learning. The citation does not predict the future and doesn't want to, but it will be undeniably true as of the casting and display. Does that suffice?

I hope we don't have to use our last few words of 70 to dilute the message. Thanks for your consideration.

-Jerome Coonen jcoonen@gmail.com +1 650-996-4738

Re: Re: Re: Advocate Decision re: this Milestone Proposal -- Bberg (talk) 12:10, 25 September 2022 (UTC)

Thank you, Amy and Jerome. As the Advocate, I have conferred with Jerome and his team about the best approach to this aspect of the citation's wording, and they have come up with new wording which should nicely address this point.

Re: Re: Re: Re: Advocate Decision re: this Milestone Proposal -- Jeromecoonen (talk) 18:41, 25 September 2022 (UTC)

Thanks again for several helpful comments. Here is a change to the last two sentences of the citation that we believe addresses several concerns.

DEPRECATED
By 1980, microprocessor companies were already implementing the proposal, which was formally approved in 1985. IEEE 754 remains the standard for reliable numerical computing.

PROPOSED
By 1980, microprocessor companies were already implementing the proposal. Once approved in 1985, IEEE 754 was widely adopted as the standard for robust numerical computing.

Length = 67 words in MS Word.

The temporal sequencing is more logical, and the stronger word "robust" eliminates a second instance of "reliable" in the citation.

Unless there are objections to this change, we will update the citation and its annotated counterpart by Tuesday.

=Re: Re: Re: Re: Re: Advocate Decision re: this Milestone Proposal -- Jeromecoonen (talk) 04:43, 28 September 2022 (UTC)=

Response to the substantive initial concern about how the citation will read in 50 years has had a ripple effect, mercifully all stylistic, not substantive. Here is the slightly modified third sentence:

PROPOSED
Its carefully-crafted arithmetic and standard data types promoted unprecedented software reliability and portability.

The difference is the modifier "carefully-crafted". The Advocate and Proposers hope to have some days of quiet, to allow further comment before the coming meeting.

=Re: Re: Re: Re: Re: Advocate Decision re: this Milestone Proposal -- Jbart64 (talk) 19:33, 27 September 2022 (UTC)=

I fully support this Milestone proposal and the final edits. Dave Bart