Talk:Pentium FDIV bug/GA1

GA Review

Article (edit | visual edit | history) · Article talk (edit | history) · Watch

Reviewer: David Eppstein (talk · contribs) 05:27, 21 April 2021 (UTC)[reply]

First reading

Overall this looks pretty good. Earwig found no significant copying. Sourcing and image licensing and captioning looks good.

The biggest problem, I think, is the entire "Microsoft Windows" subsection of "Software patches" and its two sub-sub-sections. I think the details of how to go about checking whether your 1994 hardware has this bug and which command-line flags to use to work around it if it does are arguably unnecessary detail, in violation of WP:GACR #3b, and that it would be better if the whole subsection were removed.

Another related issue (GACR #3a) is that this article is missing the boost this whole incident gave to formal verification of hardware, both at Intel [1] [2] and more widely [3] [4]

Other than that, I only found some minor copyedits to be done:

Lead:

"Professor" should be lowercase per MOS:JOBTITLES

Description:

"chips - thus": I think this should either be a spaced en-dash or an unspaced em-dash, not a spaced hyphen
"the pathological case": there is only one?

Discovery and response:

"to be of little importance and would have negligible effect": mismatched grammar

See also:

"SRT division" should be omitted per WP:SEEALSO as it is already prominently linked in the main article
I don't really see the relevance of the "microcode" link. Microcode is not mentioned here (and if it were see above) and the microcode article says that writable microcode only came into the Pentium line later.

References:

The Alexander Wolfe reference is missing information about where and when it was published, and the Cleve Moler reference is formatted to look like a blog when actually it is a reprint of a properly published source

External links:

One of the entries is a duplicate of the Moler reference

Categories:

Why "University of Lynchburg"?

I'm putting this on hold to give you time to revise the article in light of these comments. —David Eppstein (talk) 05:45, 21 April 2021 (UTC)[reply]

@David Eppstein: Thanks for picking up this review. I've massively compressed the Windows section (which was originally larger, I did consider just pulling it out - I evidently should have!) to a couple of sentences stating that Microsoft offered workarounds within Windows. I've also made the various copyedits. As for the pathological case, I am only aware of one, however it is very likely that there are others that are less 'famous' than the pair given in the article, and thus I've reworded that sentence.

I will add mention of the push for formal verification that followed this incident at some point today (hopefully!). Thanks again, ƒirefly ( t · c ) 06:41, 21 April 2021 (UTC)[reply]

@Firefly: Any progress on finding time for the formal verification issue? —David Eppstein (talk) 05:57, 26 April 2021 (UTC)[reply]

@David Eppstein: Apologies for the delay - had a crack at it, see what you think. ƒirefly ( t · c ) 17:59, 29 April 2021 (UTC)[reply]

All issues addressed, passing. —David Eppstein (talk) 07:05, 30 April 2021 (UTC)[reply]