On Monday 16 August 2010, I wrote:
> Suppose that the author decides to use g U+EE09, which has the following
> clearance matrix requirement.
>
> (0, 0)
> (2, 1)
I have now devised a collection of clearance matrix requirements for accessing
alternate glyphs for the letter g from plain text.
I have continued to use Private Use Area codes for correctness at the present
time: my hope is that character plus variation selector pairs will be encoded
in the future.
----
g U+EE04
(0, 0)
(0, 0)
----
g U+EE05
(0, 0)
(0, 1)
----
g U+EE06
(0, 0)
(1, 0)
----
g U+EE07
(0, 0)
(1, 1)
----
g U+EE08
(0, 0)
(2, 0)
----
g U+EE09
(0, 0)
(2, 1)
----
g U+EE0A
(0, 0)
(3, 0)
----
g U+EE0B
(0, 0)
(3, 2)
----
Readers interested in experimenting might like to consider the various choices
of how the codes could be used within plain text to request alternate glyphs in
the following phrase.
Eight greengage trees
Some readers might like to consider how various alternate glyphs for a letter g
could be designed and how each of them could be accessed, if at all, using the
above codes. A particular font need not necessarily have a different glyph for
each alternate that can be requested. For some fonts, two different requests
could result in the same glyph being displayed. The clearance matrix
requirements are for the boundaries of theoretical maximum boxes, the boxes
need not necessarily be totally filled to full width in any particular font.
The letter g would probably have more character plus variation selector pairs
specified than most characters, though a letter y may well have the same number
as a letter g. I have here tried to provide a good selection so that experience
of perhaps the most complicated possibility can be gained.
William Overington
21 August 2010