"perform-space-shift" should loop over all subcells. To allow this, subcells should normally be declared in sequence automatically, and the # of subcells declared should be available. Split up files better, add documentation to files. write version of Forth glossary. Update the technical reference (its supposed to be a TR!) Have Forth figure out dram-size, dram-row, and set "dcs" appropriately. implement lines and blocks in Forth need something to print all fields for a subcell (on request). how about something to check consistency of subcells? centering-hd should use Xn Yn "running out of user area" bug in Bradley Forth (associated with values?) "running out of fds" bug in Bradley Forth add zoom out capability, and generic 3d rendering fix Adler bugs word to more conveniently set frame parameters (see step-xtnd.fth). allow better bit selection and grouping in field definitions (This will let you make the 2x2.hood simpler!) lut-len make step-counts more generic set scan modes correctly move global variables into global file make byte-read double reglen Should begin-step save defaults, and end-step restore them? (rather than by "step"), then the two should be concatenated into a single list! make sure all buffers are released by "bye", and make sure that any buffers that are defined as part of the system are initialized to "unallocated" by "cold-hook". switch to more general newx (multi-level). Q: does this work if there's no *module? alloc should add to a "total-alloc" value, and free should subtract from it, so that you know how much memory you're using for step lists. VIDEO BUG: loop on sending colormap eventually causes framebuffer to drop its frame. For example, in debugging the spread mode, begin 16 0 do 1 i << not spread-mask ! colormap spread-map loop key? until eventually caused the framebuffer to lose. h# df should be a hex number, even if "df" is defined. interrupts should result in meaningful messages. print a "." each time you finish a key. add reverse to 2x2.hood life on 4x4.hood word to use next consecutive bits: 0 0 == g0 1 n == g1 1 n == g2 3 n == eighth g0 field g1 +field g2 +field ==field gh kick g0 field g1 +field g2 +field 3 x fix bug with gzip/gunzip process hanging around when abort is called