Thu 25 Apr 2013 01:33:28 PM UTC, comment #8:
I was not referring to using Adobe glyph names, which are poor in many cases, but only to using suffixes (as opposed to prefixes) for distinguishing variants of a character, as this is a very widespread convention. Probably FontForge could be changed to also allow prefixes or why not regexps.
|
Thu 25 Apr 2013 01:20:39 PM UTC, comment #7:
I regard any software using these deprecated standards in order to function as being broken. They were a poor idea to begin with, and have only caused misery since.
I just say no.
|
Thu 25 Apr 2013 01:57:32 AM UTC, comment #6:
For me the clear advantage of using the “standard” (Adobe) convention of naming variants with a suffix is that substitution tables (like onum, pnum, tnum, lnum, smcp, c2sc, sups, subs, locl, …) can be auto-filled by FontForge by specifying the suffix in the text field at the bottom of the substitution lookup dialog, and clicking the neighboring “Default Using Suffix:” button.
|
Wed 24 Apr 2013 10:51:28 AM UTC, comment #5:
Hi George,
To make it clear where I stand on this:
- consitency within the font is desirable (but a very big job
and not always clear).
- for Unicode characters, some correspondence with Unicode
names is preferable (but there are exceptions.)
- the Adobe glyph names are worthy of consideration -- only.
- how other fonts name their glyphs is immaterial.
- otherwise, the font designer must be free to name their glyphs
as they choose.
I have taken some heat, for example, for following the Unicode
names of Cyrillic letters. At one point, someone changed the
names in FreeSerif to fit their own liking (I've kind of
regretted that though...)
|
Wed 24 Apr 2013 09:10:01 AM UTC, comment #4:
No problem, I just thought it might be interesting to have glyph names consistent inside the font and with what’s found in most other fonts too.
|
Wed 24 Apr 2013 08:41:42 AM UTC, comment #3:
I agree it can be confusing, and there is no uniform system in the font.
Again, I insist that the internal names of letters are for the font designer to decide.
Glyph names really must not be interpreted by outside programs. Sure there were schemes decades ago, but these have not been maintained, have been superceded in scope by Unicode, were a bad idea to begin with, and must be discouraged and phased out.
(Yes, the scheme in PDF too -- it is a disaster that never worked right for anything beyond the most basic English text.)
I have a rough rule, for putting in new letters anyway.
Within a Unicode range, I indicate the range with a suffix:
kadeva
radeva
The suffix serves to distinguish the character from similarly-named characters in other
ranges, but is of secondary importance when working within Devanagari.
But for auxiliary glyphs having to do with devanagari, I use a prefix, because the first qustion is: what script is this glyph associated with?
Having this distinction is very helpful, for example to make it obvious in tables which is the Unicode and which is the auxiliary.
Thanks again!
|
Tue 23 Apr 2013 05:32:57 PM UTC, comment #2:
It’s better/more consistent now. But semantically, tney are still funny. Now there are variants one, two etc. of a glyph called old.
You already do have glyphs with the correct ordering at the end of the sfd file, namely C.sup, F.sup etc.
|
Tue 23 Apr 2013 10:28:38 AM UTC, comment #1:
Hi George,
I'm not against changing these names -- only your names are longer -- but more consistent with the naming of digits.
I'll think about it.
Generally, it is a mistake for any application to make use of the internal names of glyphs, and we do not support the practice. The encoding should be the sole means of accessing glyphs outside the font. (And yes I know of some very important software that uses the internal names -- to disasterous effect, quite independently of FreeFont. Old ideas die hard, no matter how bad they are.)
|
Tue 23 Apr 2013 09:22:27 AM UTC, original submission:
Glyphs like fit.1 fit.6 ss.a will be identified by some applications as variants 1 and 6 of a glyph "fit" and variant a of a glyph "ss". This might cause problems when copy-pasting from a document or when using tools like screenreaders.
The part left of the period is identified as the name of the glyph, the part right of it as the denomination of the variant.
Aforementioned glyphs should thus be named one.fit, six.fit and a.ss.
|