Re: [Devel] Mac language ID

[Antoine Leca]

Looks like I am a little late at responding, but in the case the question is
still open...

George Williams wrote on Tuesday, December 16, 2003 6:55 PM
> Slightly off topic...

Yep.

> There are 31 mac script codes (mac platform specific ids)
>    Of these 14 have documented encodings on
> http://www.unicode.org/Public/MAPPINGS/VENDORS/APPLE/

It is my understanding that they are the only that got real use. In other
words, I do not believe that any specification for Maldivian fonts and the
like were ever *thought* about...

> Some scripts have different encodings depending on the language
>    (script 0 is used for western european, icelandic, turkish, croatian,
> romanian)
>    And I think script 4 is used for arabic & farsi
>
> Does anyone have any idea what encodings are used for the scripts which
> aren't documented on the unicode site? (oriya, bengali, tamil, telugu,
> kannada, malayalam, ...)

At least for the Indian ones you cite, the idea would be a derivate of
ISCII, obviously.
Then you have to deal with all the peculiarities (ishar, bengali signs for
currencies, things like that) that were not contemplated in ISCII; this job
were done for the Northern Indian scripts, as it is reflected in Apple
published tables. I do not believe it had be done for the remaining Indian
scripts (and I believe it will never be done).

About the "...", i.e. the non-Indian ones, I would expect nothing exist.
Taking Khmer as example, one of the top expert in Burmese font encoding,
Maurice Bauhahn, is a Macintosh fanatic (no pun intented); and I never
remember reading him speaking about the Apple script Burmese encoding... And
Khmer is quite probably the most used of the remaining scripts, along with
Ethiopic and perhaps Sinhalese.

> I've asked this question of Apple and was told that there are no such
> encodings. Which I find hard to believe.

Well, I do not. After all, the list of scripts were produced one day (before
1988) to describe some part of the world writing scripts (we all know
nowadays there are more scripts) in a regular manner: so it is quite a bit
idealistic. It was then extended to deal with mere reality: adding small
differences using language ids, distinguishing "West" <<roman>> and
"Eastern" <<roman>>, Simplified Chinese, etc.
So I understand the guys at Apple did not take (lost?) the time to produce
encodings for all the scripts they described.


Going back to Freetype, it may mean that we ought to remove the scripts id
which could NOT be used for real fonts.
Since there are obviously no accepted standard beside Unicode, and
furthermore no evidence of any font using it, there are no useful ways to
deal with this piece of information. Rather, we may encounter bad use of
this, like someone trying to "escape" the Unicode standard intending to use
these ids for a proprietary encoding, and starting do give away fonts with
this, therefore impeding free communication with the rest of the community
(or worse yet, forcing people to use its products).
So it is mere bloat (of source code, however; this does not waste executable
space).

Antoine