Dear Charlie
I think it's fine to permit the new types, and thanks for the proposed text.
I would not be in favour of removing the valid_* attributes, because this is
not their sole use. It would be a backwards-incompatible change - I don't
think we have so far done that - and doesn't seem necessary to me. But perhaps
you mean to keep the attributes but deprecate their use to indicate whether a
data type is unsigned? That would be fine.
Best wishes
Jonathan
----- Forwarded message from Charlie Zender <zender at uci.edu> -----
> Date: Tue, 26 Sep 2017 13:39:38 -0700
> From: Charlie Zender <zender at uci.edu>
> To: CF Metadata Mail List <cf-metadata at cgd.ucar.edu>
> Subject: Re: [CF-metadata] Add new integer types to CF?
> User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:52.0)
> Gecko/20100101 Thunderbird/52.3.0
>
> If there is strong opposition to replacing the valid_* method in
> CF <= 1.7 with the _Unsigned method in CF 1.8 then please speak-up.
> People seem to prefer _Unsigned (as described in NUG) to the
> previous CF standard method (valid_*) as way to allow CDF1, CDF2,
> and netCDF4-classic to indicate unsigned integers in CF 1.8.
> New datasets could still use the valid_* method but that would
> restrict their CF-compliance to <= 1.7. CDF1 and CDF2 and
> netCDF4-classic datasets intended for CF 1.8 compliance would have
> to use _Unsigned instead of valid_*.
>
> This plan could be enacted with wording like this:
>
> "The netCDF data types char, byte, unsigned byte, short, unsigned
> short, int, unsigned int, int64, unsigned int64, float or real,
> and double are all acceptable. The char type is not intended for
> numeric data. One byte numeric data should be stored using the byte
> or unsigned byte data type. Unsigned types should be used for unsigned
> data if possible. If the underlying file format does not support
> unsigned types, integer or byte variables that are to be interpreted
> as unsigned must have the attribute _Unsigned = "true". The convention
> explicitly distinguishes between signed and unsigned integer types
> only where necessary. Unless otherwise noted, int is interchangeable
> with unsigned int, int64, and unsigned int64 in this convention,
> including examples and appendices. Similarly short is interchangable
> with unsigned short, and byte with unsigned byte."
>
> Comments welcome...
> --
> Charlie Zender, Earth System Sci. & Computer Sci.
> University of California, Irvine 949-891-2429 )'(
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
----- End forwarded message -----
Received on Tue Sep 26 2017 - 16:35:02 BST
