Re: Control picture glyphs

Date: Tue Aug 28 2007 - 22:17:13 CDT

  • Next message: Rick McGowan: "New Public Review Issue: #108 Ideographic Variation Database Submission"

    Quoting James Kass <>:

    > Mark Davis wrote,
    >> And the CJK VS characters should never be visible
    >> in a font that has CJK characters.
    > " a font that has CJK characters." This is an important point.
    > The developer would be well advised to include some kind of glyph
    > for VS characters, but only if the font being designed supports
    > base characters which can combine with VS characters in valid
    > sequences.
    > In other words, a font not designed to support math characters
    > and which does not support VS characters would be expected to
    > display two little boxes for a valid string consisting of a math
    > base character plus a VS character.
    > If the font supports VS characters for any other sequences, but
    > does not support math characters, and the font uses control
    > picture glyphs for VS characters, then the above string would
    > display as a little box plus the control picture for the VS. If
    > this font instead maps zero-width no-contour glyphs with VS
    > characters, then the display will just be one free standing little
    > box.

    > Because of the various possible factor combinations, the font
    > developer should retain control of font character coverage.
    > And, I just can't seem to leave this alone...
    >> And the CJK VS characters should never be visible
    >> in a font that has CJK characters.
    > ...except during input/editing, if the font designer so desires
    > and designs accordingly. In this case, though, the display of
    > an invalid or unsupported (by the font) sequence would result
    > in a display of a control picture rather than being transparent
    > to the reader. Font designers can be well aware of this, yet
    > choose to use control pictures anyway.

    I hope no one minds a rather late contribution to this very long thread.

    Here we have an interesting dilemma - namely that this would suggest a
    different display method for the text area and the input method
    display - which is really just another text area- If the input display
    for choice n is a control picture for a CJKn+VSn combination because
    it is not in the font if choosen then according to this CJKn+VSn would
    be placed in the main text and CJKn seen. Now if chioce n+1 was also a
    control character because CJKn+1+VSn+1 was also not supported be the
    font then if choosen one would see CJKn+1, which could be totally
    different to CJKn.

    The above leads to the conclusion that using the control character for
    CJK+VS any CJK+VS combination not supported is then the choice that
    works at OS level, and that this should be the default action if CJK
    VS are supported. Such a default action suggests, among other things,
    that input method editors should be CJK VS collection specific
    (including the no CJK VS option).

    Like CJK characters themselves that form the largest part of unicode,
    CJK VS will likewise form the largest part of VS, and it seems
    possible that the number of VS may well out number the CJK characters
    themselves. This may prove quite a challenge to font makers.



    This message sent through Virus Free Email

    This archive was generated by hypermail 2.1.5 : Tue Aug 28 2007 - 22:20:08 CDT