On Mon, Oct 19, 2009 at 11:06 AM, Orion Poplawski <orion@xxxxxxxxxxxxx>wrote:
> On 10/19/2009 08:36 AM, Ed Hartnett wrote:
> > <stephen.pascoe@xxxxxxxxxx> writes:
> >
> > Restricting CMIP contributors from using szip makes perfect sense. (But
> > everyone who asked me for it was from the climate community!) Yet the
> > szlib library is freely available, and license restrictions on
> > commercial writers of data will not apply to anyone in the CMIP5
> > community.
>
> Note that SZIP is not acceptable in Fedora and will not be available in
> the stock Fedora versions.
>
>
We used SZIP earlier on at GMAO in the context of HDF-4. On the positive
side, SZIP compression was much faster and most of the files were a bit
smaller. On the downside, people had difficulty reading our files. For
example, the IDL and Matlab bindings did not include SZIP support, forcing
people to do a file conversion before using our data. In the end, it was too
much trouble that we decided to abandon SZIP compression altogether. So, I
echo your concerns, using SZIP for data distribution is a bad idea. I can
see it being useful inside institutional walls for performance critical
applications. Therefore, it makes sense for the NetCDF library to have it as
an option.
Arlindo da Silva
NASA/GSFC/GMAO
