Can we recommend the use of ISO-compatible date strings, with
the caveat that time zone should always be included?
It's unfortunate that ISO defaults to local time, and that seems to be
non-negotiable.
This is what we use in the OceanSITES implementation of CF. Obviously,
it won't solve everyone's needs, but it's perfectly good for in situ data.
Regards - Nan
On 9/15/16 3:10 PM, Chris Barker wrote:
> On Thu, Sep 15, 2016 at 5:40 AM, Jim Biard wrote: > > That's a completely valid suggestion. We have to consider >
implications for backwards compatibility. Among other things, UDUNITS >
would need to change their convention, since CF follows UDUNITS in >
this matter, as with all other units. > > As long as UDUNITS accepts
ISO-compliant strings, then we're good. > And it looks like it does. > >
There's nothing that says we can't make changes. I was explaining the >
"is" as opposed to the "could be" with time units. > > The only change I
would suggest at the moment would be to "prefer" > ISO-complienat time
strings, and therefore use them in the examples. > > -CHB > > Grace and
peace, > > Jim > > > On 9/15/16 7:26 AM, Little, Chris wrote: >> Jim and
Chris B, I would like to weigh in here, please? Adhering to >> UDUnits
has merits, but once one adopts ISO8601-like notations, as >> there is
no way of specifying otherwise, people assume it *is* >> ISO8601 and
therefore a string without a time zone marker indicates >> local time
(whatever that is - Solar? Mean Solar? Sidereal? >> National legal?)
There is work just starting in the OGC/ISO >> pipeline on how to
indicate a non-Gregorian calendar using ISO8601 >> like notation, for
the WKT communities. Personally, I would >> advocate ISO8601 strict
adherence, as most of the recommended best >> practices and profiles,
such as RFC3339, are generally strict >> subsets of ISO 8601. Chris >>
>> -----Original Message----- From: CF-metadata >> Sent: Wednesday, >>
September 14, 2016 10:02 PM >> Subject: CF-metadata Digest, Vol 161,
Issue 3 >> >> Today's Topics: >> >> 1. nitpick in time axes example
(Chris Barker) 2. Re: nitpick in >> time axes example (Jim Biard) >> >>
>>
----------------------------------------------------------------------
>> >> >>
Message: 1
>> Date: Wed, 14 Sep 2016 12:34:11 -0700 >> From: Chris Barker<chris.barker at noaa.gov> >> To:
"cf-metadata at cgd.ucar.edu" >> Subject: [CF-metadata] nitpick in time
axes example >> >> I see in teh doc: >> >> Example 4.4. Time axis >> >>
double time(time) ; time:long_name = "time" ; time:units = "days >>
since 1990-1-1 0:0:0" ; >> >> IIUC, ISO 8601 requires two digits for the
time pieces [1], so that >> should be: >> >> "days since 1990-1-1
00:00:00" >> >> >> The parser I use isn't picky about this, but maybe
some are? >> >> BTW, as it's an example, we should probably throw a time
offset on >> there, too: >> >> "days since 1990-1-1 00:00:00Z" >> >> >>
or >> >> "days since 1990-1-1 00:00:00+00:00" >> >> >> [1] at least
according to wikipedia: >>
https://en.wikipedia.org/wiki/ISO_8601#Times
>> <https://en.wikipedia.org/wiki/ISO_8601#Times> >> >> >> -Chris
--
*******************************************************
* Nan Galbraith Information Systems Specialist *
* Upper Ocean Processes Group Mail Stop 29 *
* Woods Hole Oceanographic Institution *
* Woods Hole, MA 02543 (508) 289-2444 *
*******************************************************
Received on Tue Sep 20 2016 - 06:45:12 BST
