Re: geometric shapes

From: Mark Davis (
Date: Thu Mar 13 2003 - 12:39:37 EST

  • Next message: Pim Blokland: "Re: geometric shapes"

    This might be worth writing a Technical Note to start with; see

    IBM, MS 50-2/B11, 5600 Cottle Rd, SJ CA 95193
    (408) 256-3148
    fax: (408) 256-0799

    ----- Original Message -----
    From: "Frank da Cruz" <>
    To: "Pim Blokland" <>
    Cc: <>
    Sent: Thursday, March 13, 2003 08:53
    Subject: Re: geometric shapes

    > > I've got a few questions about the use of geometric shapes, like
    > > squares and such.
    > > Some of these look very similar to one another, and I don't know
    > > which ones to use in which circumstances!
    > > Are their any guidelines on their use?
    > > Just as an example, let's look at the squares. These come in four
    > > sizes: small, medium, medium small and (not specified). So my
    > > question is, as a writer: which one of these should I use when
    > > exactly? And as a font designer: what should they look like? Is a
    > > medium square (U+25FB) bigger or smaller than a square (U+25A1)? Are
    > > there any guidelines on how they should be positioned vertically,
    > > relative to normal text? Etc.
    > > The same goes for other shapes, of course. For instance, what
    > > criteria exist for, when creating a text, choosing between U+25B6,
    > > U+25B8 and U+25BA?
    > > Are there URLs available shich discuss these issues?
    > >
    > Block characters, as well as box- and line-drawing characters in general,
    > are mainly inherited from character sets in which they were included for
    > the purpose of character-cell graphics (e.g. on terminal screens,
    > text-mode DOS applications, etc), and their use makes sense only in a
    > monospace font.
    > To my knowledge, the semantics of most of these characters is not
    > anywhere, which leaves them open to misinterpretation, especially by font
    > designers who are not aware (e.g.) that such-and-such a line must extend
    > the edges of cell, so it can join up with copies of itself in adjacent
    > to form an unbroken line. Or that the two lines of "lower left box
    > must touch and bisect the top and right edges of the cell. And so on.
    > In part this is explained by the fact that the original character sets
    > which these characters were inherited were themselves documented only by
    > tables showing the glyphs, never by a description of how the glyphs were
    > be used, or line up. Only experimentation with an actual terminal
    > (e.g. VT220) or PC code page (e.g. CP437) can reveal such things.
    > The same concerns apply to some of the math characters -- integral and
    > summation sign pieces, etc, but then the character's purpose more obvious,
    > if not from its appearance, then from its name.
    > For a bit more on this topic, see the "Supplemental Terminal Graphics for
    > Unicode" writeup:
    > for Unicode 3.2:
    > - Frank

    This archive was generated by hypermail 2.1.5 : Thu Mar 13 2003 - 13:25:26 EST