William Ehrich wrote:
> Would it make sense for the code defining organizations also to define a
> bit-mapped representation, at least as an example and as a reality check?
I think, that is an excellent suggestion. In addition, the character
set standards should be extended by guidelines to font designers.
There exists a lot of application specific know-how to which font
designers have no easy access, especially concerning the various
technical and mathematical symbols. ISO 10646 for instance completely
fails to specify geometric properties of the various block graphic
characters. Their original designers had certain alignment properties
in mind how these characters can be used together to form figures,
but today's font designers are unlikely to get this right, because
ISO 10646 doesn't document these implied characteristics.
Other example: Font designers who just read ISO 10646-1 are unlikely
to understand that the glyphs LEFT FLOOR, RIGHT FLOOR, LEFT CEILING,
RIGHT CEILING are supposed to be identical with the square brackets
with just the upper or lower horizontal piece being removed, unless
they have a computer science or math degree. At least the designers
of the reference font that was used to produce ISO 10646-1 did not
understand this, these character look far from what they are supposed
to look like.
Similarly, few font designers know that it is very good practice
to give all digits the same width as the space character, because
this makes tables look much better.
It might be a worthwhile effort to write a glyph characteristics
technical report, that collects for the various fields the history
and required properies of glyphs for many Unicode characters,
especially the many technical symbols for which there exist precise
standards of how they should look like.
-- Markus G. Kuhn, Security Group, Computer Lab, Cambridge University, UK email: mkuhn at acm.org, home page: <http://www.cl.cam.ac.uk/~mgk25/>
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:40 EDT