On 3/15/2011 10:28 AM, Christopher Barker wrote:
On 3/15/11 6:02 AM, John Caron wrote:
so if currently we cant actually use months and years, because of the
way udunits handles them,
I think udunits has nothing to do with it -- one shouldn't use months
and years in the context of "unit-of-time since a date", because it's
sloppy and ill-defined. Period, end of story.
There are various date-time libraries that will let you do math like:
a_date_time + 3 months.
But in that case, the actual length of "3 months" depends on the
date-time it is being added to.
all above is correct.
but my conclusion is that its time to get a better date library.
On 3/15/11 9:13 AM, Steve Hankin wrote:
Here I think concerns of human-readable formatting and convenience are
sliding into issues of accurate encoding.
I agree -- human-readable formatting should be a very low concern for
this -- CF is about encoding data -- accuracy and clarity are far more
important.
i would say that human-readable == accuracy and clarity.
of course you also need an accurate grammer to specify legal strings,
clear semantics, and a reference library.
ncdump already offers an
option to format these values as dates. Ncgen could in principle offer
conversions to encode various human-friendly formatting options.
right -- human-readable is for utilities.
This is one of those rare areas where our responsibility as
software developers should be to push back against sloppy science, by
offering software that makes it easy to do the calculations correctly,
and help to end these sloppy practices.
Speaking as a scientist and a software developer -- hear hear!
And can we get rid of daylight saving time while we're at it? ;-)
As long as the calendar is well defined, calender dates are not sloppy,
they are the correct way to specify, uh, calendar dates.
udunit "dimensional secs since calendar date" is adequate as long as it
agrees with calendar dates. if not, it may be considered harmful.
_______________________________________________
CF-metadata mailing list
[email protected]
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
