Jim:

In our use case, software is simplified, not because of the removal of a 
multiply and an addition, but as a result of the coordinate variable values 
always having values starting at 0, and ascending monotonically and 
sequentially (i.e. 0, 1, 2, 3 …n), and just pulling specific add_offsets and 
scale_factors out of a table where the table has an entry for each type of 
product.

vey respectfully,

randy


On Apr 11, 2012, at 1:05 PM, Jim Biard wrote:

> Hi.
> 
> I'm not against expanding the use of scale and offset, per se, but the 
> reasons that you would like to use them seem (to me) to not be generally 
> applicable.  Argument 1) implies that you obtain some sort of measurable 
> benefit from saving one multiplication and one addition in your production 
> software.  If you have optimized your code so well that this represents a 
> major improvement, you are to be congratulated, but that won't be true for 
> most.  Argument 2) does not appear to be something you can count on in every 
> circumstance, nor do I think it advisable to imbue the scale_factor attribute 
> with any such interpretation as a general rule.
> 
> I am finding it often the case for swath data that coordinate variables end 
> up occupying a larger portion of a file than data variables.  I think it's a 
> great idea to address this issue.  I must admit I'm not sure what the best 
> way is.
> 
> Grace and peace,
> 
> Jim
> 
> Jim Biard
> Research Scholar
> Cooperative Institute for Climate and Satellites
> Remote Sensing and Applications Division
> National Climatic Data Center
> 151 Patton Ave, Asheville, NC 28801-5001
> 
> [email protected]
> 828-271-4900
> 
> On Apr 11, 2012, at 10:50 AM, Randy Horne wrote:
> 
>> 
>> The allure of using add_offset & scale_factor with coordinate variables goes 
>> beyond saving space.   For example, in our use case …
>> 
>> (1)
>> Makes the s/w producing the product simpler.
>> 
>> (2)
>> The scale_factor also tells you what the horizontal spatial resolution is in 
>> a representation that aligns with the coordinate space.
>> 
>> 
>> very respectfully,
>> 
>> randy
>> 
>> 
>> 
>> On Apr 10, 2012, at 3:11 PM, Etienne Tourigny wrote:
>> 
>>> If netcdf-4 is an option for you, it might be simpler to compress
>>> (deflate) the coordinate variables, this works transparently.
>>> 
>>> Etienne
>>> 
>>> On Tue, Apr 10, 2012 at 2:40 PM, David Hassell
>>> <[email protected]> wrote:
>>>> Hello Randy,
>>>> 
>>>> I, too, recently thought about this when working on some CF processing
>>>> software (cf-python), but couldn't think of a use case so didn't
>>>> pursue it - and now you have one!
>>>> 
>>>> It seems to me that the use of scale_factor and add_offset is intended
>>>> for (but not restricted to ...?) the reduction of dataset size
>>>> (chapter 8). Is the requirement in your example for space saving, or
>>>> some other convienience? I personally think that the latter case is
>>>> better served by an appropiate setting of the units attribute (such as
>>>> units="0.555555555555556 K @ 459.67" if we have converted from Kelvin
>>>> to Fahrenheit).
>>>> 
>>>> Anyway, I can't think of a reason not to allow these attributes on
>>>> coordinates - these variables can be very large, after all.
>>>> 
>>>> All the best,
>>>> 
>>>> David
>>>> 
>>>> ---- Original message from Randy Horne (01PM 05 Apr 12)
>>>> 
>>>>> Date: Thu, 5 Apr 2012 13:05:45 -0400
>>>>> From: Randy Horne <[email protected]>
>>>>> To: [email protected]
>>>>> X-Mailer: <IMail v8.21>
>>>>> Subject: [CF-metadata] Using add_offset & scale_factor with coordinate
>>>>> variables
>>>>> 
>>>>> Folks:
>>>>> 
>>>>> Appendix A, Table A.1 in the CF conventions state that add_offset and 
>>>>> scale_factor can be used with data variables, but not coordinate 
>>>>> variables.
>>>>> 
>>>>> For the data producing system I am working (GOES-R Ground Segment) it is 
>>>>> particularly convenient to make use of add_offset and scale_factor with 
>>>>> our coordinate variables.
>>>>> 
>>>>> Is there an issue with changing the conventions to allow this ?
>>>>> 
>>>>> very respectfully,
>>>>> 
>>>>> randy
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> ..............End of Message ...............................-->
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>> _______________________________________________
>>>>> CF-metadata mailing list
>>>>> [email protected]
>>>>> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>>>> 
>>>> 
>>>> --
>>>> David Hassell
>>>> National Centre for Atmospheric Science (NCAS)
>>>> Department of Meteorology, University of Reading,
>>>> Earley Gate, PO Box 243,
>>>> Reading RG6 6BB, U.K.
>>>> 
>>>> Tel   : 0118 3785613
>>>> Fax   : 0118 3788316
>>>> E-mail: [email protected]
>>>> _______________________________________________
>>>> CF-metadata mailing list
>>>> [email protected]
>>>> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>>> 
>> 
>> 
>> ____________________________________
>> 
>> Randy C. Horne ([email protected])
>> Principal Engineer, Excalibur Laboratories Inc.
>> voice & fax: (321) 952.5100
>> url: http://www.excaliburlabs.com
>> 
>> 
>> 
>> 
>> _______________________________________________
>> CF-metadata mailing list
>> [email protected]
>> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
> 
> _______________________________________________
> CF-metadata mailing list
> [email protected]
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata


____________________________________

Randy C. Horne ([email protected])
Principal Engineer, Excalibur Laboratories Inc.
voice & fax: (321) 952.5100
url: http://www.excaliburlabs.com




_______________________________________________
CF-metadata mailing list
[email protected]
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata

Reply via email to