As I understand it a grapheme is a unit in a written language, where a
glyph is the form/graph. From a dictionary: A visual representation of a
letter, character, or symbol, in a specific font and style.

A grapheme can associate with several glyphs and a glyph can represent
several graphemes.  Kind of like in "to", "too", "two" and "2" all sound
the same but have different meanings. (Actually, I can't decide if "two"
and "2" mean the same thing or not.)

I think a grapheme is like a codepoint in Unicode.

Maybe I was being a little picky to point this out, but the problem with
the 'ยต' is not a problem with the graphemes, the language units, but the
glyph representing two graphemes.

Say that a glyph that is associated with a grapheme/Unicode codepoint in
one language and another grapheme in another language, which codepoint
should be used? Or should they be combined into one codepoint? Problems I'm
sure that the Unicode Consortium worries about.

Years ago I spend days trying to find a bug in a PL/I program. I was
viewing the program on a 3278 terminal. It only displayed text in upper
case. But the keyboard entered both upper and lower case. Well, as you
probably guessed, the problem was that there was one glyph for two
graphemes.


Thanks for the tip to on using APL385 Unicode font. It worked.

And this is a case of the meaning of a word being unclear, or actually
changing meaning over time.

APL385 is actually a typeface. But as computer graphics became more
sophisticated font extended its meaning to include "typeface".

Kind of like "function" and "operator" in mathematics. Nobody really knows
what they mean.


On Sat, Apr 13, 2013 at 8:06 AM, Ian Clark <earthspo...@gmail.com> wrote:

> That's right, Don, but isn't grapheme the more general term? Unicode.org
> http://www.unicode.org/glossary/ offers "what the user thinks of as a
> character".
> So shouldn't that be "two glyphs representing one grapheme"?
>
> Different glyphs can have the same code point, eg one can be italicised.
> But in the example I gave, the two Mu's actually have different code
> points. In some fonts they've gone and used the same glyph.
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to