[UDUNITS #EYN-795297]: UDUNITS documentation

Rob,

> This does not really help. The problem is that the CF documentation insists 
> on repeating the same example of "days since ...". When another example is 
> finally shown (in the climatology section), it is invalid format (declared as 
> float, but then the data shows it in year-month-day format, complete with 
> "dashes"). It would be very helpful if the doc provided examples that work! 
> All I want to do is say something like:
> 
> date time(time) ;
> time:long_name="time" ;
> time:units="date" ;
> data:
> time="2005-10", "2005-11", "2005-12" ;
> 
> The CF and UDUNITS documentation give no clue on how to express this 
> correctly. Is "date" an acceptable datatype? If not, why are you showing 
> examples of this in the CF documentation. And what should units= be set to? 
> Or are all of these just strings?

That does seem like bad documentation, indeed.  The UDUNITS packages won't 
recognize the unit "date".

You have to realize, however, that the stand-alone UDUNITS packages were 
created independent of the netCDF packages (let alone the CF conventions).  I 
also don't interact with the CF folks all that much.

Having written that, I do know the netCDF developers quite well (the CF 
conventioners not so much).  I'll bring this to their attention immediately.  
Hopefully, you'll hear from them soon.

> -Rob
> 
> Thanks,
> Rob
> ------------------------------------
> Rob Raskin
> Group Supervisor, Science Data Engineering and Archiving
> Instrument Software and Science Data Systems Section
> Jet Propulsion Laboratory
> Pasadena, CA 91109
> (818) 354-4228
Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: EYN-795297
Department: Support UDUNITS
Priority: Normal
Status: Closed