Over the past couple of weeks I have been searching the web
  for historical and botanical references.
  
  I wrote a webpage comparing f33v with real sunflowers.  It turns out
  that f33v is only the most dubious of two "sunflowers"; the other is
  f93r.
  
  I wrote a page comparing the Vms ladies' bellies against an
  engraving by Dürer.
  
  Today I split the old Notebook-<NUM>.txt files into one file per
  day, now in Notebook/<DATE>.txt.  The occasional "summaries of work
  so far" are retained as Notebook/<DATE>-summ.txt; they should be
  merged and updated, and converted to EVA as appropriate.
  
  Jacques Guy posted on his Geocities page a large set of grayscale
  scans of herbal pages.  It occurred to me that the name of each
  plant must occur once or thrice on each herbal page, but not on
  other herbal pages.  
  
  So let's build a table of words with these characteristics.
  First, list all herbal pages:
  
    cat L16-eva/INDEX \
      | gawk \
          ' BEGIN{FS=":"} \
            ($2 ~ /herbal/){print $1;}' \
      > .her.units
    
  Now concatenate the text of those pages, and 
  remove probably insignificant details
  
    * all /q/ characters
    * map /t/ to /k/
    * map initial /y/ to /o/
    
    ??? cat `cat .her.units` \
    ???   | gawk -f eva2erg.gawk \
    ???       '
    ???   > .her-m-eva.evt
    
    foreach f ( `cat .her.dir` )
      set file = "${f:h}"
      set page = "${f:t}"
      echo "file = ${file} page = ${page}"
    end