Talk:Binary-coded decimal/Archives/2019/November

Tension over tense

David Eppstein (talk · contribs · deleted contribs · logs · filter log · block user · block log) made these changes, converting some of the descriptions from present to past tense. Edit comment: "Undo. Using present tense loses part of the point of this material, that it is about systems that nowadays are obsolete."

I reverted, with edit comment "BCD is not a (computer) system. It is a method for representing numbers, and it is still used)".

Epp re-reverted to his own text: "BCD may still be used, but the sentences in question are about its implementations on specific no-longer-used computers"

First, this is a violation of WP:BRD. Once you have been reverted you are not supposed to edit the contested material again. You are supposed to take it to the talk page.

Second, this is contradicted by MOS:TENSE:

"By default, all articles are written in the present tense, including for those covering products or works that have been discontinued."

"Generally, do not use past tense except for deceased subjects, past events, and subjects that no longer meaningfully exist as such."

"The PDP-10 is a discontinued mainframe computer family." (given as example -jeh)

I'm just not seeing any wiggle room here. The fact that System/360 and VAX computers are no longer manufactured (though some are still used, regardless of being "obsolete") does not mean the designs or their principles of operation have ceased to exist. Granted there are several articles in the computer area that get this wrong, but that doesn't mean we should ignore MOS. Jeh (talk) 20:22, 29 April 2015 (UTC)

The MOS:TENSE argument is a strong one. If I had seen that one in your edit summaries I might have been more hesitant to revert. Your weird misinterpretation of WP:3RR has nothing to do with the content and is not worth discussing. —David Eppstein (talk) 20:26, 29 April 2015 (UTC)

It's not about 3RR per se. I am aware that WP:BRD is not policy, and the "bright line" is at three reverts within 24 hours. However WP:EW is policy, and it says "Once it is clear there is a dispute, avoid relying solely on edit summaries and discuss the matter on the article's talk page". It also refers to Wikipedia:Editing_policy#Talking_and_editing, which is also policy, and says "Discussion is, however, called for if [...] someone indicates disagreement with your edit (either by reverting your edit and/or raising an issue on the talk page)." And certainly someone indicated disagreement here. Jeh (talk) 20:37, 29 April 2015 (UTC)

Re MOS:TENSE, note that past events can and should still be related in, of course, past tense. So a design decision or process that occurred in the past (even for a present-day machine!) should be described in past tense. But the resulting designs still exist in the present (even if no machines are being built from them). Jeh (talk) 20:37, 29 April 2015 (UTC)

Anyway... all those lovely packed decimal instructions are still implemented in IBM System Z architecture, which is the present-day descendant of System/360. Jeh (talk) 20:41, 29 April 2015 (UTC)

When I edit articles following MOS:TENSE I try to keep the event distinction in mind. But yes, the ideas and their descriptions still exist (present), even if no implementations exist (but most often, they do). Gah4 (talk) 21:43, 17 September 2015 (UTC)

Are hand calculators and CD players now obsolete? As far as I know, hand calculators do all arithmetic in BCD, and CDs still store the track number in BCD. (That is why 99 tracks maximum.) Gah4 (talk) 21:45, 17 September 2015 (UTC)

To keep this discussion going, note that the 8086, and still in descendants today, have BCD instructions. They just aren't as good as other machines. In 16 bit 8086, you can add and subtract binary numbers 16 bits at a time, but add and subtract BCD two decimal digits at at time, with the use of DAA and DAS. Multiply and divide one digit at a time with AAM and AAD. Also, x87 can load/store values in 18 digit BCD. Gah4 (talk) 03:11, 22 September 2016 (UTC)

(Marked claim that BCD is no longer implemented as Dubious.) Seems to me it is wrong, as I just noted. But even so, it is fairly rarely used. In years past, it was popular for COBOL compilers to use it, where little calculation was done between input and output. I don't know about COBOL compilers today. I believe it is implemented on enough machines to remove the claim. Gah4 (talk) 03:11, 22 September 2016 (UTC)