Re: netCDF Strategy
FYI, mail from Unidate as of May last year concerning libCF.
Cheers,
Nico
---------- Forwarded message ----------
From: Unidata netCDF Support <support-netcdf@unidata.ucar.edu>
Date: Mon, May 12, 2014 at 6:05 PM
Subject: [netCDF #MFY-950093]: libCF download?
To: nico.schloemer@gmail.com
Cc: support-netcdf@unidata.ucar.edu
Hi Nico,
> libCF used to be included into netCDF, but it seems that it isn't
> anymore, cf. <https://github.com/unidata/netcdf-c>.
Right, Unidata support for libCF development was halted about 2 years
ago, after the primary developer left Unidata.
> Also, the website <http://www.unidata.ucar.edu/software/libcf/>
> isn't very verbose. Is libCF available from any other location?
The most recent version, from April 2012, is still available from
ftp://ftp.unidata.ucar.edu/pub/libcf/libcf-1.0-beta10.tar.gz
The main development focus at the end had been implementing an API
for GRIDSPEC, in collaboration with a DOE-funded project, but that
library was never completed. There are currently better approaches
for conservative regridding in the ESMF regridding software:
http://www.earthsystemmodeling.org/esmf_releases/last/regridding_status.html
and its Python interface in ESMPy.
The most recent active implementation of a library for the CF
Conventions is cf-python, at
http://cfpython.bitbucket.org/
--Russ
Russ Rew UCAR Unidata Program
russ@unidata.ucar.edu http://www.unidata.ucar.edu
Ticket Details
===================
Ticket ID: MFY-950093
Department: Support netCDF
Priority: Normal
Status: Closed
On Sat, Jan 24, 2015 at 7:02 PM, Ross Gammon <ross@the-gammons.net> wrote:
> On 01/24/2015 06:08 PM, Sebastiaan Couwenberg wrote:
>
> [...]
>
>>> On 01/17/2015 01:02 AM, Sebastiaan Couwenberg wrote:
>>>> libCF is still TODO
>>
>>> ITP: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=775584
>>
>> I have basic packaging ready for libCF, but we should decide on the
>> source package name. Initially I used just 'libcf' too, but then opted
>> to change it to 'netcdf-libcf' because it's so closely related to
>> netcdf in general, and because libcf is split out from the netcdf package.
>>
>
> I prefer netcdf-libcf, keeping the package names uniform. netcdf-python
> will be the next one (the scientific python source calls its python-netcdf).
>
>> Unfortunately there are only a couple of alpha and one beta release
>> published upstream, and it doesn't look very actively developed. So I
>> have my doubts about the usefulness of keeping libCF in Debian.
>>
>
> Yes - the beta had the latest date, but it was a couple years ago. Maybe
> we should ask upstream? Packaging it and contacting them may gain it
> some traction.
>
>> One of the outstanding issues I have with the libCF packaging is that
>> it builds both libcf.so.1 and libcf_src.so.1 and the difference is not
>> clear to me. libcf in netcdf 4.1.3 only has a single version of the
>> library.
>
> We can ask about this while we are at it.
>
> Cheers,
>
> Ross
>
Reply to: