
 $VER: RGFX-Remarks.txt 4.0 (16.08.2012)
 =======================================

 Preface
 -------
 Since some time there is need for either an extension
 or a complete rework of the IFF-ILBM file format:
 one cannot store chunky graphics and storing more than
 256 colors only is possible in a very unusual manner.

 All that does slow down usage on graphics card-equipped
 systems very much.

 And IFF-ILBM's compression is not very efficient, too.

 IFF-RGB8, IFF-DEEP and IFF-YUVN aren't good alternatives,
 since they are mainly designed for 24 Bit images and/or make
 use of uncommon/undocumented compression algorithms and/or
 color spaces.

 A first step to pass this by was SView's internal
 graphics file format called "SVG" (SView Graphics),
 but it was not an IFF file format and somewhat proprietary
 in other terms, too. Additionally, there's a name-clash
 with the new vector graphics format from W3C...

 RGFX is an attempt to design a new file format that unifies
 the following enhancements over IFF-ILBM et al:

  - a real, open IFF file format
  - fully documented
  - easily extendable
  - currently supports:

 RGFX currently supports compressed and uncompressed:

  -  planar graphics                    (  <=     8 Bit integer)
  -  chunky graphics                    (  <=     8 Bit integer)
  -  15 Bit graphics as  RGB            (     3x  5 Bit integer)
  -  16 Bit graphics as ARGB            (1  + 3x  5 Bit integer)
  -  24 Bit graphics as  RGB triples    (     3x  8 Bit integer)
  -  32 Bit graphics as ARGB quadruples (8  + 3x  8 Bit integer)
  -  48 Bit graphics as  RGB triples    (     3x 16 Bit integer)
  -  64 Bit graphics as ARGB quadruples (16 + 3x 16 Bit integer)
  -  96 Bit graphics as  RGB triples    (     3x 32 Bit float  )
  - 128 Bit graphics as ARGB quadruples (32 + 3x 32 Bit float  )

 With optional PNG-style alpha channel for each 24/48/96 bit pixel
 or a single transparent color for colormapped images (upto 8 bit).
 HDR compliance via 32 bit IEEE float support.

    (* integer formats using full range (0 .. max),
       float formats normalized (0.0f .. 1.0f)

 It holds these graphics either:

  - uncompressed
  - compressed with any available XPK-Packer (deprecated)
  - compressed with libzip (LZ77/ZIP)

 This gives the following advantages:

  - you can easily (and quickly!) write raw chunky and RGB graphics
    without having to implement compression algorithms like CmpByteRun1
  - nevertheless you can use any compression algorithm you like - if there
    are really new, efficient ones available, you don't need
    to wait for an extension of the file format AND your prefered image
    processing program; no, you simply select the corresponding
    XPK packer and everything is done 'automagically'

 Remarks on compression:

  - as you may know, the XPK package now is more
    platform-independent than it was some years ago,
    so it may become available for AROS or Linux, too
  - however it subsequently became GNU software not available
    under the LGPL (but GPL) so it has been deprecated with
    Windows compatibility in mind
  - libz/zlib (ZIP/LZ77) compression is recommended now;
    it also gives a more efficient compression ratio,
    since the compressed data is not cut into single
    chunks (as with XPK)

 Other aspects of IFF-RGFX:

 - allows to store viewmode/DIPF information

 - displays really fast on RTG systems
   when using a chunky or RTG storage format.
   No more waiting for 8 or 24 Bit c2p conversion!

 - easily can be stored into and exchanged through
   the clipboard, since it's an IFF file format

 - already is supported by SView5 (on Windows and AOS/MOS)
   and some datatypes

 - and more applications...