Fraction of days, hours, minutes, seconds.
- My ISO-8601 software conversion includes fractions of days because I believe sometime in the future it will be included.ISO-8601 is unusual in that it has fraction of minutes and hours. Fraction of days is a logical extension.The question, is fractions of days time or not?If it is time, than the format should be YYYY-MM-DDT,5If not time, than the format should be YYYY-MM-DD,5 or .5My argument is that all fractions are misrepresented.YYYY-MM-DDThh:mm:ss,ssss is a misrepresentation for this reason.SS is base 60 while the ,sss is base 10. The implication of ,sss is that it is the same base as the SS seconds. So also hh,hhh that .hhh is base 24.NASA and Microsoft have fractions of time as .fff base 10.One way to rationalize time is that time has elements that are base-24 and base-60HH = base 24MM= base 60SS =base 60Therefore fraction of days is not time. Therefore the format should be:YYYY-MM-DD.ffff, much more logical to humans.Parsing the format is also many time easier.hjw
> Probably another way to solve the Leap Second issue would be to adoptWhich might be required, after all the concept "leap second" is pretty
> a new leap second scheme for decimal days, if decimal days becomes a
> common representation of parts of a day in the future.
much hour/minute/second-centric. But I doubt we could get the IERS to
also define leap-thousandths.
This sort of thing may be why ISO 8601 hesitates to support fractional
days. And why W3C shuns leap seconds.