Metrics are not critical to the programming, because the event of auto-assigning would
happen only once to a character at a given position in a string and that assignment would
not change unless the graphic designer either did so manually or reapplied the whole
process. Metrics could be permitted to vary because a graphic designer would likely look
at the result (with any character widths, any heights including ascenders and descenders,
any variations in x-height, any amount of openness of loops, and so on) and adjust as
appropriate by changing glyph assignments or by other means, just as they might adjust
leading now, and just as a secretary checks that a memo isn't one page plus only one
word on page 2. Someone could design a multiglyph font in which all glyphs for a given
Unicode code point would have, say, the same character width and the font marketer could
promote the convenience of consistency, but I doubt that that's important even as a
perceived
selling point. This is not like my spreadsheet with a randomizing function such that
whenever I open the spreadsheet it generates new numbers. Instead, a user would command
auto-assigning of glyphs to selected text or to input but, by default, once assigned no
more auto-assigning would occur, and then the user would inspect for aesthetics, fit, and
any other criteria, confident that it wouldn't change unless the user wants to change
it.
Print would be the main purpose, although the display system should make a reasonable
effort at WYSIWYG accuracy of the assigned glyphs, just as displays now distinguish
between Garamond and Times even though the distinctions are more precise in print.
Definitionally, I assume print includes things like movies shown in theatres or on
big-screen TVs, since just the scale of those displays allow more accuracy of rendering.
I don't know enough to suggest how to design this within FreeType or Fontconfig or
both or neither.
Show replies by date