A Brief Introduction to TreeRaster

[Antti-Juhani Kaijanaho <gaia@xxxxxx>]

A working tree raster appeared in GZigZag today.  New ZZ spaces will
contain four versions of this raster in the raster list which one can
browse using 'v'.

The tree raster sees the structure as a tree (a structure where one cell
is a root, and a root has many children cells, and the children may have
children of their own etc).

The structure of a tree (as defined by this raster) in the space is
the following:

  - we have a "depth" dimension and a "breadth" dimension.  Any dimension
    can be either one (this is controlled by the x and y keys in the
    default keybindings).
  
  - one of a cell's children is the cell's positive neighbour along the
    depth dimension; all other children can be reached by moving poswards
    along the breadth dimension from this cell
    
The parameter "depthhorizontal" controls how the tree is displayed.  If it
is true, then a cell's children lie on the cell's right.  Otherwise they
lie below the cell.

The key "x" controls the horizontal dimension and the key "y" controls
the vertical dimension, as usual.  Thus, if depthhorizontal is true,
"x" controls the depth dimension and "y" the breadth dimension.  The "z"
dimension is ignored.

If there are connections from the visible cells along the depth or
breadth dimensions to cells which would not be part of the tree proper
or which would normally not be shown for other reasons, the cells on the
other side of the connections are shown considerably smaller than usual,
to show that there is more there.

If the "treelines" parameter is true, then red lines show parent-child
relationships between cells.  ZZ connections are drawn in black, as usual.

Barring bugs (that you should report!), the raster will survive  and show
*any* structure you give it, however complicated.  Tuukka and I have given
the raster tough testing which it has survived.  However, the usefulness
of using this raster for arbitrary structures can be questioned :-)

Ted: this raster will not (directly) understand the structure you use
in the genealogy demo, but I believe that with a slightly different
structure it'd be possible (idea: use human children as the tree parents).
Or we could make another raster for it :-)

I'll be away (or mostly so) the next two weeks, between 2000-07-27/08-10,
inclusive.  But I'll read email.
-- 
%%% Antti-Juhani Kaijanaho % gaia@xxxxxx % http://www.iki.fi/gaia/ %%%