The current HDF-Java products can be downloaded at:
        http://www.hdfgroup.org/downloads/


HDF Libraries
=============================================================================
This release was built and tested with HDF 4.2.10 and HDF5 1.8.14. 

JDK
=============================================================================
This release was built and tested with JDK 1.7. 


Platforms
==============================================================================
This release was built and tested for the following platforms:

64-bit distribution -- runs on 64-bit systems with 64-bit JRE only
    * Linux
    * Mac Intel
    * Solaris
    * Windows

32-bit distribution -- runs on 32-bit systems with 32-bit JRE
    * Windows


Changes to the Java HDF5 Interface(JHI5)
==============================================================================
    * JAVA-1882: Changed OPEN_IDS from Vector to LinkedHashSet. Deprecated
                 int getOpenID (int index) function. Add trace logging to all
                 id create/close functions.


Major Enhancements
==============================================================================  
    * Attributes can write simple, one dimensional variable-length datatypes.


Major Bug Fixes
============================================================================== 
    * JAVA-1840: Issues with bit mask selection and usage.
    * JAVA-1837: Fixed support for scalar images and attributes.
    * JAVA-1844: Add slf4j logging document with example.
    * JAVA-1840, JAVA-1854, JAVA-1862: Issues with bit mask selection and usage.
    * JAVA-1847: Notification added to display fields that couldn't be read in 
    			 datasets.
    * JAVA-1666, JAVA-1878, JAVA-1880: Issues with filters and dynamic loading.
    * JAVA-183:  Correctly show bitfield and opaque datatypes.
    * JAVA-1855: Object and Dataset Region references can be selected and viewed.
    * JAVA-1853: Fix issue with any stride above 2 causing an off-by-one-error 
                 and not displaying the last row/column.
    * JAVA-1863: If attribute is of Integer class, set the correct type and size.
    * Improved internal management of compound datatypes, limitations still apply.
    * Improved usage of user.dir and user.home properties as defaults to hdfview.root
                 and new property hdfview.workdir properties. This also fixed
                 issues with the location for the open file dialog.


Limitations / Known Problems
==============================================================================  
    * HDFView updates attribute changes directly to the file. However, the
      metadata window will only update the values of attributes when focus is 
      changed in the treeview. The metadata window will only show the name
      change when the file is reloaded.
    * HDFView does not support undo operations; the editing results go to
      files and cannot be recovered.
    * HDFView does not support conversion between HDF4 and HDF5.
    * HDFView supports only indexed images with RGB color model or 24-bit
      true color images.
    * Directly using multiple dimensional arrays causes a performance
      problem because there is no direct mapping for multiple dimensional
      arrays between Java and C.
    * HDF 4 APIs with function pointers are not supported in HDF-Java.
    * HDF 5 APIs with function pointers are selectively supported in HDF-Java.

    * Specific to HDF4:
        -- Deleting an object or attribute from an HDF4 file is not
           supported. Deleting objects in HDF4 may leave dangling objects
           or corrupt the file.
        -- HDFView cannot modify compressed HDF4 SDS if the rank is greater
           than two because the existing compression algorithms supported
           by HDF do not allow partial modification to a compressed
           data stream.
        -- HDFView cannot modify HDF4 vdata. HDF4 only writes vdata by
           records.
           HDF-Java reads data by vdata fields because HDF-Java does not
           support the data structure of a vdata record.

    * Specific to HDF5:
        -- Writing compound data is limited to simple cases. For example,
           base compound fields that have primitive types such as integers,
           floats, or characters can be written. HDFView does not write
           complex compound data. For example, HDFView cannot write a dataset
           that has a compound datatype with a region reference.
        -- Writing variable-length data is not supported except for datasets
           where each data point is a single variable-length string or a
           one dimensional integer type.
        -- There is no single JNI function that can handle HDF5 library
           functions that have different return types for version 1.6 and
           version 1.8. An additional function has been added to the JNI.
           For example,
             * public synchronized static native
                      int H5Tget_size(int type_id)
             * public synchronized static native
                      long H5Tget_size_long(int type_id)


Other Notes
==============================================================================