# File TODO
# RasMol2, Roger Sayle
# Glaxo Research and Development
# Last Modified 24/10/94

# Current Projects:

[1] POVRay output file format.
------------------------------
    Implement a "write povray" command, potentially based upon
    Eric Suchanek's "pdb2pov" program much like "Personal 
    Computer-based Visualization of Molecular Models by
    Available Ray-tracing Software", Arnold S. Dion, J. Mol. Graph.,
    Vol.12, March 1994, p.41-44.

[2] Improve method used to atomic determine bonding.
----------------------------------------------------
    Create a table of bonding radii for each element type. Modify
    DetermineBonds in molecule.c to accept a boolean parameter to
    specify whether a fixed radii is used or the bonding table.
    By default use fixed method for molecules with >255 atoms and
    the accurate method for smaller molecules. Add a command 
    "CONNECT" to force re-calculation of connectivity with a
    boolean parameter

[3] Stereo Left/Right [Hardware?] Images.
-----------------------------------------
    By using a data structure to describe the current frame buffer
    containing information such as width, height, bytes per line 
    and a pointer to the buffer, it should be possible to implement
    different eye views to either side-by-side in a single image
    of two separate images. The latter can then be used to support
    the "SGIFullScreenStereo" X extension.

[4] Add Gausian and MOPAC format files.
---------------------------------------
    This requires a "torsional driver" to handle atom co-ordinates
    in "internal" z-matrix format. Clever parsing of these files
    avoids the need for babel's internal/cartesian distinction.

[5] Add support for thick ribbons.
----------------------------------
    Add the command "set ribbons " to specify the depth
    of the ribbon strand. Currently solid ribbons have zero depth.
    This should also improve the gouraud shading of ribbons.

[6] Write out solid ribbons and strands in vector PostScript.
-------------------------------------------------------------
    Create a structure to store the complete knot description at
    every alpha carbon. Depth sort these knots as part of the
    painter's algorithm. Write them out in the PostScript using
    suitable macros. Need to handle oblique polygons specially,
    and may need to subdivide to more accuracy than when rendering
    on the screen.

[7] Improve Cylinder Quality.
-----------------------------
    The current algorithm for rendering cylinders for stick
    bonds and backbones is slower and less efficient than it
    could be. This also results in strange aliasing effect 
    for oblique cylinders. Correcting these problems should
    permit cylinders of much large diameter, such as cylinder
    representations of helices. Similarly add the ability to
    have flat (and rounded) cylinder ends.