Follow-up Comment #1, bug #66903 (group groff): For context, * AT&T Version 1 to AT&T Version 6 UNIX used mm/dd/yy * AT&T Version 7 UNIX used mm/dd/yy in nroff output and Month dd, year in troff output * 2BSD used mm/dd/yyyy * 3BSD to 4.1BSD were inconsistent * 4.2BSD used dd Month yyyy (no comma) * Since 4.3BSD-Tahoe (June 1988), BSD has been using Month dd, yyyy (as introduced in v7 troff) consistently for more than 35 years, and this is documented in mdoc(7). * The ISO 8601 variant yyyy-mm-dd is common in older GNU manuals, though i don't know where it first appeared in manuals. It certainly has no tradition in UNIX, neither in the AT&T nor in the BSD branch.
I think using yyyy-mm-dd is a bad idea. Even though endorsed by ISO, purely numerical mm-dd and dd-mm are notorious for being easily confused with each other. The traditional AT&T UNIX and BSD format Month dd, yyyy is preferable because it is unambiguous and reads well for human beings, which is the audience of manual pages. Being easy to parse for a computer, on the other hand, is not a goal for manual page output. There recebtly was some discussion on the tech at mandoc mailing list, for example https://marc.info/?l=mandoc-tech&m=174172793214691 Existing mdoc(7) manuals tend to be highly consistent in this respect, so attempting to change the recommendation for mdoc(7) is likely to make matters worse even though causing huge effort in at least four major projects. Changing the recommendation for man(7) would have less adverse impact because man(7) manuals tend to be less consistent in this respect. Adjusting the recommendation for man(7) to match the firmly established one for mdoc(7) seems like the best option to me in the long run. _______________________________________________________ Reply to this item at: <https://savannah.gnu.org/bugs/?66903> _______________________________________________ Message sent via Savannah https://savannah.gnu.org/
signature.asc
Description: PGP signature
