- there aren’t very many accented glyphs
    D in particular å
- “6” and “a” have a higher baseline than “e”, “o”, and “8”
D there is no ’ or ‘ (or “”)
    - they could still use some work; ” is lighter than “ by too
      much. But they are quite distinct from ".
    - all of '`‘’ are too short, I think
- still need to redraw everything but lowercase with a better line-width:
  the lettering in general is a little too heavy, and I think I should
  switch to stress at 1:00 and 7:00, and a little less contrast;
  instead of 120, 10:1, 45°, maybe 100, 6:1 (0.167), 60° from the horizontal
    D trying this on the digits. It does seem to be working okay so far...
    D next should be '`"[]()
        - maybe skew the [] a bit? “[bd]” looks a little odd.
    D then /
    - then TAIS (most common word-initial letters)
    - then :;,. (,. may already be okay)
    - then OWHBC
    - then <=->
    - then PMFDR
    - then ELNGYU
    - then VJKQZX
    - then #$%&*+
    - then \?^_
    - then {|}~
    - then dotless i
- "h" is still too vertical and should be rotated or skewed a bit to
  the right
- “h” seems to reach down too far, as do “m”, “r”, and “b”
- “r” is too tall
- “p” tilts too far to the right; “n” and “t” not far
  enough. Maybe reserve some of the next box to the right for
  upper-right-hand bits?
- in "FB" the "F" is noticeably shorter than the "B"
- things like <-> and <=> don’t line up properly
- “teach” has too much space between “ea”
- "y" isn’t diagonal enough
- the up-arrow is on the | instead of the ^
- "gn" in "assigning" is too tight
- in "subsequent" the "q" is too low by comparison to the letters nearby
- also, "q" could stand to be redrawn with a horizontal top, a la Monaco
- redraw lower case with new line weight etc.:
    - then etao
    - then insrh
    - then ldcum
    - then fpwygb
    - then vkxjqz

- some thoughts on absolute sizes of things:
    - a 1024x600 window is 127x35 at the default font size of about
      8x17px and 255x66 at the minimal readable size, which I guess is
      about 4x9px.
          - 8px for 500 units is 62.5 units per px or about 21 units
            per color channel. Consequently, stems must be at least 21
            units apart to be distinguished at all, at least 42 units
            apart to be distinguished reliably, and at least 62.5
            units wide and apart to be colorless; horizontal lines
            must be at least 62.5 units apart to be distinguished, and
            125 units to be distinguished reliably.  Character width
            variation must be at least 11 units to get even a single
            third-of-a-pixel of variation.  The stem of the "t" is
            about 40 units wide; similarly for the stem of the "h",
            most of the bowl of the "e", and so on.  Thin pen strokes
            get down to about 14 units; the thickest in most glyphs
            are about 50.  127 times a third of a pixel of character
            width variation could add up to 40 pixels, which is 5
            characters. Acceptable as long as characters like space
            and -= (and things that similarly occur in great numbers)
            aren’t abnormal.
          - At 4px, you have 125 units per px or 41 units per color
            channel, so you need stems to be 82 units apart to be
            distinguished reliably, and 125 units wide and apart to be
            colorless. Character width variation is a lost cause at
            that size.

Several hypotheses:
- Stroke width contrast makes for better readability.
- The ideal angle for a wide-nibbed pen on an LCD is closer to the
  vertical than 45°.
- The descender space is inadequately used in today’s standard roman
  fonts, including the fixed-width ones.  See the font on
  <http://www.topatoco.com/merchant.mvc?Screen=PROD&Store_Code=TO&Product_Code=QC-PAP&Category_Code=QC> for another example of m and n descenders.
- Monospace character cells don’t need to be rectangular.
02:10 < KragenSitaker> my underlying theory here is that western typography 
                       since ancient Rome has been unduly Platonist
02:10 < KragenSitaker> concerned with geometric forms and timeless perfection
02:11 < KragenSitaker> calligraphic forms like kaisho represent an opposing 
                       aesthetic sensibility, a wabi-sabi sensibility rooted in 
                       impermanence and imperfection
02:13 < KragenSitaker> the popularity of Papyrus seems to me like a cry out for 
                       more wabi-sabi in Western typograph.
02:13 < KragenSitaker> y.
02:14 < KragenSitaker> (And its unpopularity among Designers seems to fit that 
                       theory. it's a very non-Platonist, non-modernist font.)
02:14 < KragenSitaker> but Papyrus is a display font; although it's gorgeous, 
                       it would be hard to read a book set in it, I think
02:15 < KragenSitaker> so that's the underlying theory. unfortunately at this 
                       point I'm a pretty terrible typographer, so even if the 
                       theory is good, I may not be able to tell