[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[UDUNITS #ONU-754617]: Comments on udunits-2

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: [UDUNITS #ONU-754617]: Comments on udunits-2
  • Date: Mon, 10 Sep 2007 12:47:20 -0600
Harvey,

> http://rain.aos.wisc.edu/%7Egpetty/physunits.html
> http://xml.coverpages.org/OlkenMeasurementUnitsSyntax.html
> http://en.wikipedia.org/wiki/Level_of_measurement

Thanks for the references.  It looks like Fortran-95 can do some
pretty cool stuff.

> Petty's paper explains the idea of non-integer powers of base-units.
> Note that I am NOT suggesting allowing unit strings with non-integer
> powers.  I assume they are useful only during calculation, not for
> input/output.  Thus ut_format could simply round powers to the nearest
> integer.

Like I sand in my previous reply, I'd rather restrict argument to
integers where appropriate.

> One feature of udunits-1 which I disliked was the merging of the Julian
> and Gregorian calendars (with a specific calendar change date).  The
> date of conversion varied from country to country.  In fact some
> countries (e.g. Russia) did not change to Gregorian until the 20th
> century.  Gregorian dates should be treated as a separate unit from
> Julian dates and it should be possible to convert from one to the other.
> I note your comment in the manual about now having second thoughts about
> the wisdom of supporting dates at all.

Yup.  I hate the time-offset unit.  I should never have created it.
Anyone who uses it as a replacement for a full calendaring system
(which is what people try to do) has rocks for brains.  The only
solution would be to implement a full calendaring subsystem with
the UDUNITS package, and I'm just not willing to expend that much
effort on so little a benefit.

> Note Olken's distinction between 'measures' and 'coordinates'. Udunits
> has the concept of a time-origin (often related to dates), but this
> 'coordinate' concept is relevant to non-time dimensions. E.g. eastings
> and northings are relative to an origin.  Note that it is legal to
> divide one measure by another, but not one coordinate by another.  In
> other words (see wikipedia URL above) a measure is an 'ratio
> measurement' while a coordinate is an 'interval measurement'. I'm not
> sure what all this means for udunits but maybe any unit could have an
> origin.

Most units can be made relative to an origin via the ut_offset() function.


Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: ONU-754617
Department: Support UDUNITS
Priority: Normal
Status: On Hold