\	$Id: TODO,v 1.1.1.1 2009-06-09 18:28:50 anton Exp $	

See also DEVELOPMENT VERSION-README and BUGS.
At the time being, many TODO items are in file NEXT.


make reproduce spot configurable, i.e. free-neighbour-spot?

check all c-l and l-s for being world dimensions instead:		done
mark when scrolling is a deal				DADA MARK: scrolling
	c-l 	0 (dim-spots) !
	l-s 1- 	1 (dim-spots) !

simple-expressions-all and the like should have sublists

GUI type user								done

update README

rectangle selection in world view.					done

range selection in scan displays	define-bar-range		done

all-lists get re-built. Should be implemented with sub-lists.

floating point in word-usage.fs  dfVARIABLE  fVARIABLE
try to fix CREATE (without DOES>)
use (don't-count-words)
make better user interface
try smarter CREATE counting mechanism storing current

test on other screen sizes

test and run benchmarks

brew is absolutely stupidly relying on a  9 16 fixed font size. 	done
does it need to know just now?
Mouse position could be stored on clicking for cases you would need pixels.

df-function-number and df-function-within for floats		done
	
use new scan.fs							not yet

continuous display type hits					not yet
based on count on everybody

new nuc-function-scan step display type				done

step display adaption: step-display-adapt-size			done

mouse support							NEED HELP

scrolling in menus on different screen sizes			done

fix ?step-display-sanity	fit screen size			done

BUG when mouse outside configured screen size, bottom line	fixed

File selection should display files ordered, directories first.   done

Fileselect: use  (start-char)				no, use scroll-to-line

Use 'compile-switches.fs'					not yet

Extend menu elite.

Reporting in elitism mode.					done

named indices for (scan-detail-index) and the like		no, use xt's
  * nuc vars
  * world vars
so: (scan-detail-index) should become (scan-detail-xt)		planed
  * gives an address	bitmap / int / dfloat
  * gives a value	bitmap / int / dfloat

scan scores and other functions on nucs or spots		done

fitness-list							score-list done

world:step:spot							done

recording and logging of world name changes.			done

saving of all worlds.						hmm...

Record file start and worlds list.				done

clone world							done

check compile options, i.e. no dfloats

file-structures on compiled genes				done

check code-file-masks, if they are still in use			done
same for run-mode

check READMEs							done

maybe-run-benchmark.fs						done

remove genes/ directory		CVS

factor out individuals						partly done

left over tmp files				PENDING

show-coloured-on-nuc-var-eq					obsolete?
show-coloured-on-nuc-var-diff					obsolete?

real count zero in average and display		PENDING


re-think 't' and 'o' key bindings in brew main screen

more generic foreground/background colouring functions		PENDING
same for dfloats						PENDING
float>colour functionality					PENDING

dfloats for world-spots-localized.fs			delayed until used

OFFSET: use au's

\ unique: placeholder

float>buffer-string and nan inf...

maybe split checksums in an integer and a float part?

world directions

\ cells work on *present* food plane, 'world-do' does copy it to the future.
\ WHY?

nuc menu: diversification flags on one line

pad cp for speed...

consistent use of upper case stack type symbols as addresses of the
lower case type.

\dfVARIABLE

use diff-blocks everywhere

\ ***********************  nuc pointers:  ************************ 

rename 'menu-same-key-entry' to 'menu-key-same-entry'.


multiple sequential gene 'chromosomes'
each one get's mutated separately having it's own mutation-rate, gene pool,
mutation types...

tune open-allocate default size

make population control world local
copy population control parameters (e.a.) between worlds

clone world

recursive page see

fg-colour-field							fg: done
bg-colour-field							bg: not yet

scrolling and shifting visibility window

test diff-genome-pool when more possibilities are used

Rewrite probability lists to be more like normal lists?		hmm?
Lists using negative offsets for internal data?

Nucs from individuals menu should get a new genome id when sown or set !
(Fixing that might break benchmark compatibility, test first...)
Maybe individuals should always get a new genome id?			done
|
Related bug:
Changing genome manually must change genome id!				done
|
Maybe store a genome-id for each genome? ('noop' would be zero).

obsolete:				delayed (compatibility)
do-on-world-field
world-condition-color-map
world-condition-color-diff


HELP menu

fix COUNTING-WORDS		partly done (CREATE DOES> not always working)

extraction mutation type.

string interpretation on function key

'2-ascii-scale' rename and maybe reorganise

move world variables when entering a world.

selecting spots in higher dimensions.

world list.

save-world-variables

all visible worlds (as a new visibility level).
maybe starting only on second dimension?

(big-bang) parameters the other way round.

text-display on info line

test and usage example in 'menu.fs' using up to date syntax...
include 'menu-leave' in there.

give compile time hints
alignment
unused cells
...

future-change-individal
does not make much sense any more with new-worlds		see below ;-)
food>future same.
yes, it does, because I run 'localise-spot-data' FALSE now...	ok

<?diversify-planes> is defined in brew.fs but depends on new-worlds.
Must be tested anyway...

rethink spot-vars@ spot-vars!

Reactivate 'nucs-checksum' when bigFORTH 386-Linux rev. 2.0.2 is fixed.
workaround installed, reactivated.

save-nuc when nuc is on trial
save-nucs-genes-trial

.keybindings needs an update.					partly done

move benchmarks/ to INPUT/

'living' and 'die'...

show coloured on nuc flags: 'show-coloured-on-nuc-flags-NAND' ...

save and load nucs from individuals menu.

men-step-presets

eliminate diversify?-A ...

is-gene?        \ dummy FALSE at the moment
see-genome      ." NOT DONE YET"

finish make-ekey-map.fs

do i want <diversify> ?

A-quality>color-scale


\ there's a bunch of files and features which are not used yet:

\ these for the planed gene conditionals generation  [ ... no I didn't use it ]

\ longstanding:
mouse.fs		\ dummy based on old Atari version


Code mutation currently only work on *top* level of the genome.

respect max-stack-effect

more random functions to choose one

\ not implemented yet...
mouse interface

\ not implemented yet...
undo

look-at1
look-at2
...
