[Oberon] Hacking S3 fonts

[eas-lab at absamail.co.za]

I have not yet looked at how the data structure(s) of *.Scn.Fnt
relate to the font set.  Except to note that OB2ULINE.Scn.Fnt
has extra chars added to Oberon10.Scn.Fnt , and the size is
increased from 2303 to 2777.
And Shanghai.Scn.Fnt = 21944 bytes; has big arty pictures - try it !

Perhaps those who know the data structure(s) could see that my
idea is not viable, and save me looking deeper.

1. Is this the correct understanding of a glyph: 
 my present font shows "o" and "l" as an oval and a vertical line;
by repositioning these 'glyphs', we can build the chars: "pbqd" ?

2. this leads to the main question.
Here's an answer I got:
>  No none of the Indian typewriters type a single character and
> then step like in English typewriter.

So apparently the (all of the many) indian script(s) are 'built up'
by 'overwriting glyphs', before moving on to the the next char.

In principle do you see any problem in using the existing system to:
- construct an 'alphabet' of glyphs = set-A;
- extend the existing system to be able to merge the bit-patterns 
      corrresponding to elements of set-A, and place the merged 
      bit pattern in a temporary *.Scn.Fnt .
- extend this further to be able to reposition any glyph/bit-pattern
     before merging it.
    ??

The idea would be to be able to (eg.) construct "E" from 2 glyphs
and some 'positioning commands'.

I'm guessing that this is how indian typewriters work.
If this is feasible, I want to take a laptop to india and let some
school kids construct their native script for n-o.
Initialy I'm interested in Malayalam.

Does anybody see problems with this idea - before I start.

Chris Glur.
--