A GNU COBOL status update

If it makes you feel better, I thought it was a reference to Y2K, and all the hemming and hawing about COBOL programs not handling time (and say, quirky things like DST) very well.

And I heard tell that Napoleon tried to decimalise time. That 10-day week *really* didn't work out well.

At my workplace (and I think it's EU rules) many people are simply NOT ALLOWED to work more than six consecutive days, on safety grounds.

Cheers,
Wol

Redefine day..

That would be unfair: some people would always have to celebrate their birthday on a weekday, while others could always celebrate on a Sunday ;-)
(replace weekday and Sunday by the days of your liking)

There have actually been multiple proposed calendar reforms that would've eliminated the changing days of the week, usually by introducing an intercalary day (i.e. a day that is not counted as any day of the week). Of those, the World Calendar probably got the closest to being implemented, but various religious groups opposed it (on the grounds of "my holy day falls every seven days, and fiddling with the calendar would mess that up").

If it had been implemented, the World Calendar would have introduced a three-day weekend beginning on Saturday, December 30* and ending on Sunday, January 1, with the middle day being called "Worldsday" and having no day of the week (or month, for that matter). In leap years, another intercalary day is inserted between Saturday, June 30, and Sunday, July 1. The calendar has the same number of days as the Gregorian calendar and uses the same leap year formula, so there would be no need to add or subtract days in order to switch to it (as was done with the Gregorian reform).

If you want something more standard and widely-supported, you could use the ISO 8601 week-numbering year, which always has a whole number of weeks, but usually does not agree with the Gregorian calendar as to the nth day of the year. The week-numbering year works by labeling whichever week contains the first Thursday in January (according to the Gregorian calendar) as "week 1," and then numbering weeks consecutively until you get to "week 1" of the following year. This means that, for example, if a year begins on Wednesday (according to the Gregorian Calendar), then the week-numbering calendar says it begins on December 30 (the preceding Monday). If a year begins on Monday, then both calendars agree on the beginning of the year, and the nth day of the year, but they will disagree as to when it ends. Because the week-numbering year defines its starting point relative to the Gregorian calendar, it cannot fall too far "out of sync," and periodically makes up for the lost days by having 53 weeks in a year. No individual year is closely aligned with the tropical year, but in the long run, the discrepancies cancel out.

Unfortunately, there's probably no good way to combine these types of reform: You can either have a year that follows the tropical year, as the Gregorian and World calendars are, or you can have a whole number of weeks, as the ISO week-numbering year does, but you can't do both.

* The calendar shortens December to 30 days, as part of a broader reform intended to produce more consistent month durations and four equal quarters.