IDC's versus Egyptian format controls

[James Kass via Unicode]

Richard Wordingham wrote:

> There is another possible use of the latitude given by TUS 5.0 to 10.0
> and possibly earlier.  I can certainly imagine a case where someone
> writes a font so that an unencoded character may be manipulated like any
> other character.  He has two choices - he can put it in the PUA, or he
> can make it the ligature for the IDS.  If he chooses the former, and
> then the text and font are separated, the recipient of the text is left
> with tofu for the character.  If he chooses the latter, the recipient of
> the text would at least have the IDS.  I think the latter outcome is
> the better outcome.

Yes, I think it's much better to leave the unencoded ideograph
unmapped (not assigned within the font to a Unicode code point) and
treated as a font ligature.  If the unencoded ideograph is encoded,
then the ligature glyph would be mapped to the actual character, of
course.

When estimating the complexity of the look-up tables involved, please
keep in mind that, as the complexity of the ideograph increases, so do
the number of different ways of breaking down that ideograph.  And all
of those ways would need to be accomodated in the look-up tables.

For example, U+2A7FF "��", according to my notes, can be described in
two pieces (⿰卒巽).  The right half "巽" can be further broken down into
three components (⿱⿰巳巳共).  The left half could also be broken down
further.