Re: [netcdfgroup] taking szip write capability out of netCDF snapshot code...

All,

Adding a quick comment to support the position that Ed and John are taking: given the restrictions on szip support I feel pretty confident that the CF group would elect to forbid its use in CF-compliant files. A key reason for enthusiasm about netCDF-CF is the degree to which it promotes interoperability. Files that can be read only by a restricted (licensed?!) set of software at certain sites would not meet this test. Quite the opposite; we'd be introducing a serious stumbling block to interoperability. The need for client writers to support multiple client software builds would be a HUGE downside to offset the potential advantages of szip.

   - Steve

=============================

John Caron wrote:
I think its a mistake to promote the use of a proprietary compression scheme, unless it was much better, which i doubt it is. If the szip patent holders would like to release their IP claims then that would be another matter. I also think its fatal that theres no java implementation, and theres doubt whether one could legally be written. We would then be in the position of writing netcdf-4 files (by the C netcdf library) that cant be read by the java netcdf library. OTOH, trying to read existing szipped data is a good thing if possible, since there may be Important Stuff in There, that wants to be set free.


Ted Mansell wrote:
Ed, et al.,

Szip Cons:

* No existing Java version.
* License restrictions for commercial writers of data.
* Some (or most) netCDF-4 installations will not be able to read szipped
 files without rebuilding netCDF.
* Will not (and should not) be used by CMIP5 effort and (probably) other
 important archives.
* Due to licensing szip will not be available in stock Fedora
distribution. Fedora is a very popular Linux distribution, and at least
 some other free software distributions will probably feel the same
 about szip licensing problems.



All of these apply equally well to hdf5, so why is it a big deal? szip is not a default option for hdf5, so probably most folks don't use it, anyway. I think if you already put it in, it might as well stay in for those who want to use it and don't need to worry about compatibility. (I'm sure plenty of commercial software doesn't even read netcdf4 yet, either.)

As for Fedora users, they should be able to install stuff that they want. I'm not complaining that OS X doesn't have szip in the stock install.... get the tar file and go.

If licensing is really an issue, it could be a separate config option to enable writing of szip (e.g., --enable-write-szip), and let --with-szip just enable reading.

Is it just a matter of time before the Java version catches up? Not that I use it....

Best,

-- Ted

_______________________________________________
netcdfgroup mailing list
netcdfgroup@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit: http://www.unidata.ucar.edu/mailing_lists/

_______________________________________________
netcdfgroup mailing list
netcdfgroup@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit: http://www.unidata.ucar.edu/mailing_lists/



  • 2009 messages navigation, sorted by:
    1. Thread
    2. Subject
    3. Author
    4. Date
    5. ↑ Table Of Contents
  • Search the netcdfgroup archives: