The R Project SVN R

Note: Like the TASKS file, this is way out of date, hopefully replaced
      by a bug/enhancement requests system Real Soon Now (TM). -pd


                THE R PROJECT LIST

              Dream, dream, dream ...
                - The Everly Brothers


Here is a list of projects which we think could enhance R substantially.
Some of these we are thinking about doing ourselves, but others could be
taken on by anyone.


 o  WINDOWS 3/95/NT PORT.
    Robert has this in hand and is nearly done.


 o  POWER MACINTOSH PORT.
    This is temporarily on hold, pending funding or Ross having some
    free time.  [ We have to do this.  Our teaching labs are Mac based. ]


 o  LIBRARY STRUCTURE
    We need a way to "attach" libraries.  This should happen in
    conjunction with "save" and "load", and with some sort of hashing
    to speedup lookups in libraries.


 o  3D GRAPHICS
    We need something like "persp" for plotting surfaces, although it
    would be nice to have something more sophisticated.  Interesting
    gadgets to look at include:

    *  "Lab-3d" has some pretty nifty techniques for hidden line
        removal, but it looks like is done in "image" coordinates.
        This would be a problem for any device independent system.

    *  "zplot" has a very nice user interface and provides a method
       of viewing surfaces in 3d.  An efficient set of hidden line
       and hidden surface routines.

    *  "plotmtv" is more static, but has some interesting contouring
       ideas.


 o  DYNAMIC GRAPHICS
    Statistics for the video game generation ...  We need to move to
    an event-based model for this.  See the 3D graphics above plus:
 
    *  X3d looks like it would be worth a look.  It provides very
       fast response for complex wireframe structures.

    *  XWireframe has some nice little features.

    *  XGobi is nice too.


 o  MULTIPLE ACTIVE DEVICE DRIVERS.
    The basics of this is actually pretty simple to implement.  The
    hard part is thinking about how to manage display lists.  It would
    be nice to have a choice of "in-memory" or "disk-based" display-list
    storage.


 o  MATHEMATICS IN PLOT LABELS
    We have an expression parser and can grab parse trees.  All we
    need is access to font metrics for bounding boxes (done now) and
    some time to plow through a couple of appendices of the TeX Book.
    [ Ross and Paul Murrell are working on this ... ]


 o  LINEAR MODELLING
    The whole "glm" suite needs to be rethought and redone.
    We need "aov" and friends (pretty easy given what we have).
    Labelling for nested specifications needs to be handled correctly.
    Additional functionality (drop1, update ...).


 o  NONLINEAR MODELLING
    Was that Doug Bates I heard rustling about over there in the
    corner ... :-)


 o  GAM MODELS
    The basic code is at STATLIB.  Did Trevor Hastie put out an
    update to this?


 o  LOESS MODELS
    The code is at NETLIB.


 o  RECURSIVE PARTITIONING
    There is some very nice looking code by Terry Therneau at STATLIB.


 o  TIME SERIES
    A state-space/structural models package would be nice.
    We have a partial spectrum analysis package.  It needs to be
    multivariate though.


 o  MAPPING
    We have the Becker/Wilks sources for constructing map data bases.
    All we really need is the drawing code.