Glyph Name problems

levonk's picture

Hi,

I am designing Armenian fonts. The glyph names in the unicode standard are in the form "ARMENIAN CAPITAL LETTER AYB".

Initially I only used the last word, which is the name of the character in Armenian, and it was working fine, except it did not recognize otf fonts as being Armenian fonts.

I changed the glyph names to the standard name, and now it gives me errors when I try to generate fonts. Plus I do not know how to use these names in the OpenType features.

Are the names genreating the problems? Is there some way of addressing these glyphs?

charles ellertson's picture

In cases like this, where there is no "Adobe name," I have always used the Unicode index as the name -- e.g.,

uni0531

for the name, and

0531

for the Unicode index.

FWIW,

Charles

levonk's picture

So, there should not be any problem if I use my own names. That is, programs and the OS do not check for names it is only for internal use.

charles ellertson's picture

No, that is not correct -- at least, if you want *transportablility* of the document. I believe Adobe products do use names, Microsoft programs use the Unicode index. But you (we) need someone like Adam or Thomas for the definitive answer.

In the meantime, try

http://groups.msn.com/fontlab/tipsandtricks.msnw?action=get_message&mview=0&ID_Message=3065

Charles

twardoch's picture

> So, there should not be any problem if
> I use my own names. That is, programs
> and the OS do not check for names it
> is only for internal use.

If you use your own names, the users who use your fonts in PDF documents will not be able to perform searching or copy-paste of text under some circumstances.

It is very advisable to use the Adobe/FontLab glyph naming guidelines for all fonts:
http://groups.msn.com/fontlab/tipsandtricks.msnw?action=get_message&mview=0&ID_Message=3065

Regards,
Adam Twardoch
Fontlab Ltd.

levonk's picture

Adam, I read the "Glyph Naming and Encoding" article and now I am more confused. According to the article a glyph name should not contain spaces, which is good and solves my problem. However, the glyph names defined by the unicode standard 4.0 for the armenian characters include spaces.

Is this a problem with the unicode standard? Can I ignore the unicode standard naming scheme to have a more compatible and reliable one? Is it possible to ask that they change the naming?

ps. the site http://partners.adobe.com/public/developer/en/opentype/glyphlist.txt has the name "Aybarmenian" instead of "ARMENIAN CAPITAL LETTER AYB". Should I use that name?

hrant's picture

Levon, parev! Nice to have a fellow Armenian (and a Beirutsi at that) on this forum.

I wish I could help you to any useful extent, but my only experience with making Armenian Unicode fonts is from the Fontographer days - I haven't had a chance to get to grips with how FontLab does it, although soon enough I will have to bite the bullet. I expect it's a lot easier than what I had to do with Fontographer though (I remember having to run it through Michael Everson for a quick technical quality check). So I think those who have more Unicode experience (even if it's not Armenian-specific) will end up helping you more.

BTW, I saw your other thread too, but I have dozens of Typophile threads flagged for eventual :-/ perusal so it's hard to find the time for everything. Plus I owe Miguel (Sousa) a full personal reply on his project first... I hope one day soon I can help, somehow.

hhp

twardoch's picture

Levon,

1. The human-readable long names used in the Unicode Standard have *nothing* to do with glyph names.

2. The article http://groups.msn.com/fontlab/tipsandtricks.msnw?action=get_message&mview=0&ID_Message=3065 points to an old location of the Adobe Glyph List for New Fonts. Unfortunately, the Adobe site incorrectly forwards the URL to http://partners.adobe.com/public/developer/en/opentype/glyphlist.txt

This URL should *not* be used for devising glyph names. The correct URL for the Adobe Glyph List for New Fonts is http://partners.adobe.com/public/developer/en/opentype/aglfn13.txt

We will soon post the corrected guidelines.

Since the Armenian glyph names are not in the AGLFN, please use the uniXXXX names. So for ARMENIAN CAPITAL LETTER AYB, you should use the name "uni0531", just like Charles explained.

Regards,
Adam Twardoch
Fontlab Ltd.

Thomas Phinney's picture

I'll get that URL linking fixed.

And, as Adam says, the Unicode standard does not have glyph names. It has character names, which are a different thing entirely.

Cheers,

T

levonk's picture

Thank you guys.

Is it possible to suggest the glyph names to adobe, or whoever is responsible? There are two different dialects (Eastern and Western) with their own pronunciation and spelling, but it is possible to find a middle ground.

levonk's picture

Hrant, verchabes kezme badaskhan me.
Iv'e been doing extensive research on the subject of glyph names and the only armenians I find are Eastern armenians that don't seem eager enough to help. It is nice to finally find someone that speaks the same language.

hrant's picture

> verchabes kezme badaskhan me.

:->
Sori, khenti bes sbaghvadz em vertcherus.

Mintchayt, asor masin lour ounis?
http://typographicbeirut.lau.edu.lb/
Yes hon bidi ullam.

--

As a rule, "official" efforts tend to follow the standards (e.g. pronunciation) of the political state (if it exists). So as much as it's awkward for you and I, we basically have to follow that lead here.

The "middle ground" you speak of is something that comes naturally to us Armenians :-) but it sort of goes against the cultural ideology at the heart of something like the Unicode Consortium. I myself apply such a middle ground (for example when informally transliterating proper names, for the benefit of non-Armenians) but I don't expect that of formal institutions (at least not Western ones, not these days).

hhp

Thomas Phinney's picture

The problem is, on those rare occasions when glyph names matter, you need applications to recognize those glyph names. There is no technical advantage to be gained from adding new glyph names, only incompatibilities.

This is also why the list of glyph names acceptable to be used for new fonts is much much shorter than the list of glyph names we suggest applications try to recognize. In the first case, we are recommending that font developers use the narrowest, most compatible set of names and naming schemes. In the latter case, we are recommending that application developers try to recognize glyph names from any of a variety of competing standards, including some that conflict with each other (like Adobe's and Apple's for some names).

Currently, we have no intention of ever expanding either of those lists again.

Regards,

T

twardoch's picture

> Is it possible to suggest the glyph names
> to adobe, or whoever is responsible?

Levon,

there is no need for that. The current naming guidelines practically cover all cases, and most importantly, they set up a referential scheme to an internationally acknowledged standard (Unicode/ISO 10646). I realize that the names "uniXXXX" are not necessarily human-readable so using them in production may be cumbersome. This is why some developers are using their own "human-readable" glyph names in production and rename the glyphs to Adobe/FontLab-compatible names before shipping their fonts.

For batch glyph renaming, you can use the macro available from
http://steroids.fontlab.net/

Regards,
Adam Twardoch

evertype's picture

Personally I am hoping that I can find some way of using FontLab; next week I will be at the Unicode conference in Berlin, and I will take the tutorial. I have a lot of legacy stuff with legacy names, and have had pretty much terrible problems dealing with encoding, names, and so on in FontLab. There's got to be a solution, but it's not out of the box.

levonk's picture

Using the uniXXXX names is not a big problem. However, I would prefer "human-readable" names. It will reduce the time I spend on preparing kerning tables, OT features, etc.

Adam, I will try the macro you suggested. If I can work with my glyph names to prepare everything, and then change them back to the uniXXXX name, it will save me a lot of time, confusion, and errors. But it means that I will need to keep 2 copies of each font.

John Hudson's picture

Levon, the way I handle this is to make two FontLab encoding files, one with 'development name' (i.e. human friendly) and one with the glyphs names that will be written to the final font (uniXXXX, AGL, or some combination). Then I use Adam's macro to switch glyph naming back and forth as I need to. Generally, I do all my work with the development name, and only switch to the final names when I need to generate a font. I only maintain one FontLab source, usually with the development names.

Syndicate content Syndicate content