Changes between Version 33 and Version 34 of NetCDF_Improvements

Timestamp:

Dec 23, 2011, 6:12:10 PM (12 years ago)

Author:

etourigny

Comment:

--

NetCDF_Improvements

@@ -23 +23 @@
 === Issues with the current implementation ===
 
- * NetCDF created by GDAL not CF-1.0 compliant for geographical grids (issue #2129) - fixed in trunk
+ * ~~NetCDF created by GDAL not CF-1.0 compliant for geographical grids (issue #2129)~~
  * netcdf driver does not read and write CF-1.0 coordinate system properly for projected CRS (issue #2893) - mostly fixed in trunk
-   * support for read and write of projected coordinate values - fixed
-   * export std_parallel variables into a single standard_parallel netcdf attribute - fixed
-   * add proper datum import and export support - fixed
+   * ~~support for read and write of projected coordinate values~~
+   * ~~export std_parallel variables into a single standard_parallel netcdf attribute~~
+   * ~~add proper datum import and export support~~
    * issues with specific projections (see wiki:NetCDF_ProjectionTestingStatus and generic report in #2893): 
      * rotated_pole (#4285 and #4251)
-     * Stereographic vs. Oblique_Stereographic (#4267)
-     * LCC1SP - no scale_factor in CF-1 (#3324)
+     * ~~Stereographic vs. Oblique_Stereographic (#4267)~~
+     * ~~LCC1SP - no scale_factor in CF-1 (#3324)~~
      * CEA : how to deal with scale_factor_at_projection_origin vs. of standard_parallel1 
-     * PS: how to deal with standard_parallel vs. scale_factor_at_projection_origin
- * add CF compliance checker to the autotests (issue #4233) - fixed
- * support "coordinates" attribute for projections which are not supported by GDAL/PROJ.4 (i.e rotated pole) (issues #4285 and #4251)
+     * ~~PS: how to deal with standard_parallel vs. scale_factor_at_projection_origin~~
+ * ~~add CF compliance checker to the autotests (issue #4233)~~
+ * support "coordinates" attribute (and associated curvilinear grid) for projections which are not supported by GDAL/PROJ.4 (i.e rotated pole) (issues #4285 and #4251)
 
 
@@ -44 +44 @@
 === Issues with the current implementation ===
  * The import process does not use the Conventions attribute and does not consider different possible Conventions (COARDS, CF-1.0, CF-1.5, etc.)
- * netcdf history global attribute should be updated. Does GDAL offer a history mechanism? - needs new function in gdal.h (issue #4297)
+ * ~~netcdf history global attribute should be updated. no mecanism for saving commands in gdal, due to security issues (issue #4297)~~
+ * implement setmetadata and setmetadataitem, tricky because other nc attributes are already written to file
 
 === Issues that have been resolved (in trunk)  ===
@@ -74 +75 @@
  * Support Create() method so that gdalwarp can be used on a netcdf file - #4221 (committed to trunk)
  * ~~Fix access to different netcdf file types: nc (classic), nc2 (64-bit) and nc4 (NetCDF4 with HDF5 support) - #3890 and #2379~~
- * Add option for file type (nc,nc2,nc4) at export (#3166) -> fixed in trunk, but needs testing in Win32
- * Support compressed netcdf (DEFLATE with NC4 or NC4C) (#4404, #4221)   -> fixed in trunk
- * fix the y-axis for import and export (issues #4251 and #4284) - in trunk option for bottom-up
- * support packed data (with add_offset and scale_factor) transparently (see bug #3797) - will not implement as no sufficient need
- * support for signed/unsigned Byte data - #4221 (committed to trunk)
+ * ~~Add option for file type (nc,nc2,nc4) at export (#3166)~~
+ * ~~Support compressed netcdf (DEFLATE with NC4 or NC4C) (#4404, #4221)~~
+ * ~~fix the y-axis for import and export (issues #4251 and #4284)~~
+ * ~~support packed data (with add_offset and scale_factor) transparently (see bug #3797) - will not implement~~
+ * ~~support for signed/unsigned Byte data - #4221~~
 
 References:
@@ -85 +86 @@
  * http://www.unidata.ucar.edu/software/netcdf/docs/BestPractices.html#Packed%20Data%20Values
 
-== Various bugs that need fixing ==
+== Various bugs / wishlist / questions ==
 
- * support import of unequally-spaced grids
- * support of lon/lat grids from CF grids which are not supported
  * can we define dim vars before band vars, without performance hit? nc__enddef??
  * should we set NO_FILL for performance gain? nc_set_fill(ncid, NC_NOFILL, &old_fill_mode);
- * implement setmetadata and setmetadataitem, tricky because other nc attributes are already written to file
 
 == Back-porting to 1.8 branch ==