[Oberon] Re. What does Char(hexA7) look like ?

shark at gulfnet.sd64.bc.ca shark at gulfnet.sd64.bc.ca
Fri Dec 5 23:13:46 CET 2003


Chris,

pe> > When Desktops.OpenDoc encounters § it shows 
> the "section sign" glyph. 

cg> I don't see how this is possible.
> I think the charcter is rendered/shown by the particular
> *.Scn.Fnt 

Software can use any number of fonts to render  
a text to an image.  Also, associated with each 
font can be an encoding vector which allows a 
permutation of the order of the glyphs in the 
font.  Also, a font can have fewer characters 
than the character set in use.  These concepts 
were present years ago in the use of PostScript;
probably even earlier in use of HP PCL & etc.

What happens when Desktops goes to render an
HTML text to a viewer?  Most characters are simply
represented by their ASCII ordinals and the 
glyphs come straight from the *.Scn.Fnt as you
say.  Yet when Desktops finds § it knows to
get the section sign from some other font.  
Similarly for several other "above 127" characters.
Apparently Desktops can not recognize a name such 
as § and renders just the name itself.
This is something a volunteer might fix.

A text editor is simpler than Desktops.  Edit and
ET recognize nothing in § beyond a string of 
six characters.  If *.Scn.Fnt is an ISO 8859-1 font, 
then A7(F) should render a section sign.  
Conversely, if *.Scn.Fnt is meant to be only an 
ASCII font then characters above 127 are not 
specified.  There are two ways to fix this.  One 
is two extend *.Scn.Fnt to 256 characters.  The
other is to give Edit or ET the capability of 
invoking appropriate fonts for characters above
127 as in Desktops.  The first approach is simple 
and tidy.  The second makes use of existing fonts, 
avoids duplicating glyphs across fonts and requires 
more cpu operations to execute.

These "hasles" relate to the creation of Unicode.
Have a look at EditTools.Tool.  EditTools.*Unicode 
might handle the "above 127" characters but it 
doesn't appear to.

Well, no shortage of programming work to be done.

Regards,        Peter E.

       http://carnot.pathology.ubc.ca/




More information about the Oberon mailing list