RE: N3412: Last Resort Pictures

From: Philippe Verdy (
Date: Tue Apr 01 2008 - 16:39:47 CST

  • Next message: "Re: N3412: Last Resort Pictures"


    Are the letters chosen within the glyph mandarory? I don't think so for such
    But the choice of letter A in Latin is not very helpful (note that the tiny
    comments inserted on the black rounded square box will be unreadable in most
    cases, so only the central glyph will remain identifiable).

    I'd much prefer to see a Latin letter R, which is distinctive of the Latin
    script when compared to Basic Cyrillic, Greek and Coptic (however there's
    serious doubt that basic latin will not be supported, so having to display
    the lst-resort glyph for characters in the ASCII block is very unlikely; it
    is less unlikely for characters in the Latin 1 block (but if needed, the
    distinctive character of this block should probably be Latin letter Thorn,
    or Latin small letter y with diaeresis). The choice of Omega for Greek is
    perfect, as well as the chice of letter IA for Cyrillic, and letter IE for
    extended Cyrillic.


    Also I'm wondering about the need for last-resort glyphs for the surrogate
    blocks (proposed U+E0297?-U+E0299?). Any renderer that would render those
    glyphs would not comply to Unicode if it displays them without parsing pairs
    as a single codepoint, and then rendering characters, and not code points.
    If a document contains isolated or unpaired surrogates, it does not conform
    to Unicode for plain text.

    So may be its inly use will be for some "visible controls" mode in editors
    when all characters in the supplementary planes will be decomposed. But this
    mode would not be very helpful as all supplementary characters, including
    those supportd in fonts, would be decomposed.
    * If some of them are recognized, or the blocks in which they are referenced
    is recognized, then the last-rest picture for each recognized supplementary
    block will be more helpful.
    * If this is for characters that are in still unsupported supplementary
    blocks, may be these characters will be shown as pairs of last-resort
    glyphs, but my opinion is that the renderer would more probably render some
    better last-rest glyph, such as a reversed question mark in a dotted square,
    for ALL characters within undefined or unsupported blocks.
    * No decomposition into visible surrogates should be made visible to users.
    If a document is not conforming to Unicode, every occurrence of an isolated
    or unpaired surrogate should be marked distinctly in the text editor using a
    glyph that will look more like the one you used in your proposal for illegal
    non-characters (i.e. with a black background and the forbidden symbol), and
    it should be clearly distinct from the last-resort glyphs used for valid
    (assigned or unassigned) character blocks. For me, surrogates are NOT
    characters, so if a glyph is ever needed for them, it should clearly be
    similar to those used for other NON-CHARACTERS.

    > -----Message d'origine-----
    > De :
    > [] De la part de Michael Everson
    > Envoyé : mardi 1 avril 2008 22:04
    > À : Unicode Discussion
    > Objet : N3412: Last Resort Pictures
    > Please see

    This archive was generated by hypermail 2.1.5 : Tue Apr 01 2008 - 16:50:44 CST