[mpeg-OTspec] Re: [OpenType] MS Proposal for a new Name Table ID

[Peter Constable]

Probably what really should be used are not BCP 47 _language tags_ but rather BCP 47 _extended language ranges_. For instance, “*-Latn” is an extended language range that can be used to indicate applicability to content in any language written in Latin script.

In Windows 7, I implemented a feature that’s reflected in the Windows 7 Fonts control panel that used metadata tags to capture a description of what languages a font was designed for – i.e., would be a value add for users of that language. In that metadata, we used tags following BCP 47 syntax but with a small variation: the script subtag, and only the script subtag, was mandatory. That achieved the desired result; BCP 47 extended language ranges provide for the same capabilities but using something defined in an open industry spec.

Btw, I think I managed to convince Elika Etemad last year that the CSS spec for lang pseudo-class selectors should support extended language ranges. I haven’t followed up to see if that’s getting clarified in a draft revision to CSS selectors.


Peter

From: mpeg-OTspec at yahoogroups.com [mailto:mpeg-OTspec at yahoogroups.com] On Behalf Of John Hudson
Sent: Tuesday, January 08, 2013 2:47 PM
To: opentype-migration-list at indx.co.uk; mpeg-OTspec at yahoogroups.com
Subject: Re: [mpeg-OTspec] Re: [OpenType] MS Proposal for a new Name Table ID



I'm beginning to think that two levels of information might beneficially
be indicated using the kind of tagging in this proposal. I can think of
plenty of fonts in which the information of interest is script rather
than language, e.g. an Indic font that contains a Latin 8-bit subset for
purely technical reasons and that is not intended to be used to set
Latin-script text. In that case, one would want to be able to indicate
that e.g. the Devanagari script is the intended use, not particular
languages using that script. The latter might be secondary information
if, e.g. a font is particularly intended for Marathi and not suitable
for Hindi.

As I understand it, the BCP 47 language tags include provision for
indicating script, but only as a secondary indicator applied to a
language tag, e.g.

zh-Hant (Chinese written in Traditional Chinese script)
zh-Hans (Chinese written in Simplified Chinese script)

And for font tagging purposes this is backwards: the first -- and often
only -- indicator of intention needed is a script tag. This suggests to
me a hierarchy of script and language.

Now, this might suggest something like Adam's proposed use of existing
OTL tags to indicate intention, only in this case the script and
language system tags. I don't think this works though, because I
regularly include Latin kerning and other features appropriate to the
character set in fonts that happen to include a Latin subset, even
though the font is not intended for setting Latin-script text, and hence
have a OTL feature tree in the font. After all, even if a font is
not intended for setting text in a Latin-script language, the occasional
untransliterated English word may occur within e.g. Hindi text, and
should be properly displayed with appropriate kerning. Also, the OTL
language system tags indicate something other than language (what I
would call -- to use terminology in a way similar to Martin --
particular writing systems whose characteristic differ from the defaults
of an individual font's behaviour). And, of course, the OTL language
system tags, other than , indicate exceptions rather than intentions.

So I think a tagging system to indicate intentions is a good idea, but
think it needs to provide for both script and language. It should
definitely be possible to tag a script or scripts as intended use,
without needing to tag any languages. I'm leaning towards requesting
some kind of hierarchical model, because I can imagine wanting to
indicate something like this:

Latin
Cyrillic
- Bosnian

Of course, Bosnian can be written in the Latin script too, but a font
wouldn't necessarily need to indicate specific support for this, since
the required characters and glyphs wouldn't differ from norms for any
other languages, while the Cyrillic requirements might.

JH

--

Tiro Typeworks www.tiro.com<http://www.tiro.com>
Gulf Islands, BC tiro at tiro.com<mailto:tiro%40tiro.com>

The criminologist's definition of 'public order
crimes' comes perilously close to the historian's
description of 'working-class leisure-time activity.'
- Sidney Harring, _Policing a Class Society_

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.aau.at/pipermail/mpeg-otspec/attachments/20130109/5e2235f4/attachment.html>