ID for language-invariant strings
petercon at microsoft.com
Thu Mar 20 04:23:13 CET 2008
> From: ietf-languages-bounces at alvestrand.no [mailto:ietf-languages-
> bounces at alvestrand.no] On Behalf Of Tracey, Niall
> I'm going to nail my colours to the ZXX mast.
> The developer is entitled to choose any identifier he likes when naming
> his font. It could be "Gothic Cursive Heavy", it could be
> "GothCursHeav", it could be "GCH123", it could be "&*hhdjg%&&^" or it
> could be "1000100101110010010101".
In the application I have in mind, a reference form for a font's family name, I think developers are very unlikely to use something like "&*hhdjg%&&^" or "1000100101110010010101", or even "GCH123". Font developers give their fonts real linguistic family names (to the extent that names generally are linguistic).
Now, when it comes to a font's unique font ID string (OpenType name ID 3), then I agree that something like "1.011;ADBE;ArnoPro-Bold" should be tagged "zxx". And a font's Postscript name (e.g. "AJensonPro-LtItDisp") probably should be tagged "zxx", although the OpenType spec already has explicit requirements regarding the use of platform-specific language IDs for English.
When it comes to family names (OT name ID 1 or name ID 16 or name ID 21), though, these are generally linguistic. They may have some abbreviations, as in "Arno Pro Smbd SmText", or some non-linguistic qualifiers, as in "Blackletter 686 BT" ("BT" = Bitstream is a linguistic abbreviation, but "686" is non-linguistic). And there are the rare exceptions: in a sample of over 12000 fonts I'm looking at, there are about 95 (< 0.8%) that have purely-numeric family names. Generally, though, the names are linguistic.
More information about the Ietf-languages