Some software applications, such as version 9 of the desktop publishing package made by Serif, can only access 256 character codes.
This is a situation which has been around for some years and is because characters are each stored using one 8-bit byte. Many modern applications are implemented with characters each stored using two 8-bit bytes.
In the older systems, there was often a need to use some characters from Unicode-based fonts, where characters use codepoints each stored using two 8-bit bytes.
This was, and still is, achieved using codepages, whereby the 128 codes from $80 to $FF are each mapped to a Unicode character codepoint of two bytes. That is, although the software package using the codepage can still only access a total of 256 characters, of those 256 characters 128 can be any Unicode character.
In Unicode, $80 to $9F are for control codes and $A0 to $FF are mostly for the accented characters used in the languages of Western Europe, langauges such as French, German, Portuguese and so on.
There are various codepages which can be used for various purposes, such as for scripts from the Indian subcontinent and so on.
One particular codepage, often the default codepage on computers in Western Europe, in fact has the characters from $A0 to $FF the same as in Unicode, so although the codepage structure is in principle looking them up and replacing one glyph by another, it is actually replacing a glyph by a copy of itself, so it seems as if no look-up is being done.
However, the codes from $80 to $9F are used for being replaced by other Unicode codepoints.
It is possible that the software application which you are using is using that codepage. If so, in oreder to produce a font, you need to find out a list with two columns in it, the first column having 32 codes from $80 to $9F and the second column having thirty-two Unicode codepoints. The one corresponding to $80 is U+20AC, as Erwin mentioned earlier. So, the next stage is to find those other thirty-one Unicode codepoints. Then, when making your font using Font-Creator, you need to put your required glyph designs at the thirty-two Unicode codepoints in the second column of the list.
However, there could be some problems in doing that as four of them are used for smart quotes and some programs will substitute those for the ’ and " characters. That is, if one types “hello” into the program, it may change the first " to become a 66 style open quote and the second " to become a 99 style close quote. However, not all packages do that, so it might not happen.
I note that you wrote “I have changed symbols for codes from $80 to $C2 in my font.”.
If you are not using $C3 to $FF, a better solution to the problem may perhaps be not to use $80 to $C2 but to use $A1 to $E3 instead, that would avoid the problems of the smart quotes.
However, proceeding with the original problem as presented, the answer to how to find the thirty-two codes for the second column of the list is on the following page.
http://en.wikipedia.org/wiki/Windows-1252
Some would be easier to use than others, for example OE oe and the S and Z accented characters. Some such as the smart quotes could lead to problems.
Anyway, I hope that this helps.
William Overington
30 March 2007
and you run complete composites on that glyph, it will attempt to create the symbol from the capital “T” and capital “M”. If those glyphs don’t exist in your font, the result will be to erase the existing 