#4 trace-back

[Piotr Sawuk]

there are several problems with the trace-back function
(when used without setting marks):

1) when altering the bit-field after running some
steps of the animation the changes get lost after
the second step backwards due to recalculating
the whole thing from start.
2) which does cause severe inconsistency when
editing step 0.
3) and which also causes step 0 to be displayed
as empty when it happens to require re-calculation.
(i.e. go back from step 2 to step 1 so that no such
calculation gets done, and then further back to
step 0 does produce an empty field even though
all the data is still there)

I'm thinking of a real backwards-tracing, which would
solve all those problems, but which would require major
change of interface. as it is this program treats animation
as if the game of life would be a linear film, but for real
research it would be useful to calculate all possibilities
of previous steps upon request, and navigate through
them in some tree-structure (with the most advanced
point of time as the stem of the tree). this would also
mean the end of the "step x" dogma since people would
wish to see what happened before step 0. maybe a
button for going back the same way one did get there
could still be useful, but a tree of possibilities should
be the way to go.