From: John Hudson (john@tiro.ca)
Date: Fri Aug 24 2007 - 18:09:23 CDT
Andrew West wrote:
> So I would say that a font that does not support variation sequences
> should follow Mark's suggestion and map all VS characters to an
> invisible glyph, whereas a font (such as Code2000) that does support
> variation sequences should give VS characters a visible glyph, which
> can be available for use in "show hidden" mode, but which will be
> stripped out by the rendering system in normal use.
> But what a font should not do, in my opinion, is simply leave VS
> characters to be rendered with the .notdef glyph.
Ideally not, but there are tens of thousands of fonts out there in the wild, most will
never be updated, and if they are updated only a fraction of users will end up with the
updated versions. And what about when Unicode encodes new control characters that you
might also want to be handled in this way? Then all the fonts will be 'broken' again.
This is an issue that cannot be practically addressed at the font level. On the other
hand, applications could relatively easily suppress display of the .notdef glyph when the
backing store includes and unsupported control character that is better not displayed. So
I would look to application developers to solve this particular display issue.
Fonts are very good at displaying what the makers of the fonts want them to display. It
shouldn't be up to a font to not display something that should not be displayed. The
.notdef mechanism itself is primarily an application operation only supported by the
inclusion of a dedicated glyph in a font. Modification of that operation to suppress the
.notdef glyph in certain circumstances should also be an application operation.
John Hudson
-- Tiro Typeworks www.tiro.com Gulf Islands, BC tiro@tiro.com Do not begin to paddle unless you intend always to paddle. - St Jean de Brébeuf, instructions for missionaries, 1637
This archive was generated by hypermail 2.1.5 : Fri Aug 24 2007 - 18:10:51 CDT