(See the Parser file.) NextWord roughly returns parse-->(w*2-1) (but it worries a bit about commas and full stops).

[ WordAddress w; return buffer + parse->(w*4+1); ];
[ WordLength w; return parse->(w*4); ];