## IDEAS FOR THE PEP/NOM SYSTEM

 A document of ideas of the pep/nom system that may never be implemented. Often
 I find that new commands aren't necessary after all.

 - add an 'eprint' command that prints the workspace to 
   stderr not stdout. This can be used for warnings etc.
 - substitute a [latex] ampersand for a normal one
 - some randomisation for nom. 🙵  ampersand
 - archaic words
 - 'getx;' command that appends the pep://tapecell to the 
   pep://workspace the number of times that is in the pep://accumulator.
   This can be sort of emulated with eg:
    * emulation of getx;
    -----+
     count; replace "2" "..";
     replace "3" "...";
     replace "4" "....";
     replace "." "abc ";
    ,,,,
 - sublists starting with items --
 - empty item

### UNICODE AND READING 

  Unicode doesn't actually define what is a 'character', instead it uses code
  points and combining marks (diacritics, underlines, etc). So the question of
  what the nom://read command should do is a bit tricky. Should it read a byte,
  or a (utf8) code point or a grapheme cluster (the series of codepoints and
  combining marks that makes up a single visible character) ? Usually the
  answer is the last: A grapheme cluster, but not always.

  So the nom://read command could have several modes, and it could 
  even read the data as a byte and treat it as 'binary'.