This is not intended as a complete TODO list, but it does represent
a good place to start.

In the short term, certain key user interface functionality must
still be revived or newly developed, such as:

* convert all window drawing to a direct Draw() method, instead of
  relying on updateGL() and paintGL() (cannot depend on rigorous
  synchrony between window updates and simulator time-steps otherwise);
  main window is already converted
* make windows record their position and size changes to the settings
  file when they are manually moved or resized
* record all (or a specifiable subset of) Polyworld windows, using the
  new PwRecordMovie/PwMoviePlayer tools (they used to all be embedded
  in a single effective window--the entire Iris screen--but are now
  multiple windows that need to be managed separately and kept in sync)
* make dump files work again (to save state and restart the program) 
* write status file again
* make the main view's camera rotate again 
* get the overhead view working again 
* get zoom working in the overhead view 
* allow overhead critter following (and toggle on/off, and set fitness rank)
* get the genetic-spread graphics working again
* make a click on a critter highlight its POV sub-window (new) 
* make a click on a POV sub-window highlight the critter (new)
* make clicks in either place select that critter for brain and overhead
  monitoring (new)

However, the core, scientific simulation engine is fully functional,
I believe.  Longer term goals involve the use of information-theoretic
measures of complexity to assess the neural network dynamics, and,
in turn, to use such measures of complexity as a fitness function
to drive evolution more rapidly (than pure natural selection) towards
neural systems with greater effective intelligence.  Besides measuring
and using complexity in this formal sense, the complexity, now in an
informal sense, of the environment itself will need to be increased
by means such as:

* make food grow in an organic fashion, instead of randomly 
* allow organisms to pick up and put down pieces of food 
* allow organisms to pick up and put down pieces of barrier 
* introduce multiple food types and let an organism's genes
  determine how much it needs of each type 
* introduce geographic trade-offs between energy consumption
  and energy availability

And so on, with all such extensions geared towards providing more
available niches to foster speciation or to provide mechanisms that
enhance the evolutionary usefulness of collaboration.

-------------------------------------------------------------------

Recently done:

* fix bug in PwMoviePlayer that results in disembodied, unpopulated little
  MenuBar floating at the top-left of the window (no idea where that's
  coming from currently)
* get the digital movie recording and playback revived (from rev20 or rev21) 
* brain map window won't close when close button is clicked
  (selecting Hide Window from the Window menu works fine)
* fixed colors of population graphs
* fixed bug in smite option (had to be updated on critter death, with
  new on-demand behavior)
* add "smite" option so that when the world fills up (to its maximum
  number of critters), the least fit (that are over 1/2 their maximum
  lifespan) can be killed to make room for organisms that want to
  reproduce 5/31/05
* put the Xcode shell script phase into its own target, so it completes
  and generates new moc files (or not, as needed) before the dependencies
  are analyzed for the main Polyworld target 5/31/05
* add code to dump neural net anatomical and functional data for
  network analysis using Olaf Sporns's tools
* fixed a couple massive memory leaks (all known memory leaks are fixed!)
* made it so the gene separation isn't calculated unless it's needed
* made it so the miscegenation function is easily turned off
  (by setting misccritters to -1 in the worldfile--that's the way it's
  now checked in)
* added frame rates (instantaneous, recent, overall) in both frames/second
  and seconds/frame to the text status window
* get the brain monitor window actually working
* get the strip of vision pixels back in the brain monitor window 
* fix the ChartWindow's behavior - results do not seem to be averaged into
  the first half after the second half fills, and the second half is not
  cleared before points start being plotted into it again (after filling)
