[Debian-in-workers] Pending Changes todo for Package Maintainers

[Kess Vargavind]

2011/11/13 Mahesh T. Pai <paivakil at gmail.com>:
> Right now, a fight is going on in the Unicode's Indic list with one
> guy on one side and rest of the people familiar with Unicode on the
> other - the issue being why Assamese uses the Bangla script.

Yes, and it has been suggested by the list administrator that the
issue is closed. [1]

In short the English name for the script is Bengali (or Bangla), the
Bengali name is বাংলা লিপি (bān̐glā lipi) and the Assamese name is
অসমীয়া লিপি (asamīẏā lipi); I’m using a common transliteration scheme
for both languages. Simply different languages, nothing else to it. It
is like saying that both the English and Finnish alphabets are using,
or are based on, the Latin script. But in the case of Assamese and
Bengali it would be Brahmi rather than Latin.

[1] http://www.unicode.org/mail-arch/unicode-ml/y2011-m09/0148.html

> As I gather from the unicode mailing list, one or two characters
> exclusive to Assamese are included in the Bangla / Bengali code block.

I am only aware of the following character being proposed for
addition, but it is not really relevant for this discussion, I think:

  U+0980 BENGALI ANJI [2]

As well as several proposed additions and corrections to the character
descriptions in this block to better reflect the characters’ usage,
see for instance [3].

[2] http://std.dkuug.dk/JTC1/SC2/WG2/docs/n4157.pdf
[3] http://std.dkuug.dk/JTC1/SC2/WG2/docs/n4148.pdf

> Coming to topic of this thread, if the fonts look different, what we
> have is a difference of style; as in "Roman vs. Gothic vs. sans
> seriv".

I would not say that they differ in styles. More that they differ as
much/little as, say, Spanish and Danish writing.

> But since there are one or two codes specific to the Assamese language
> in the Bangla block, I suspect that these exclusive-to-Assamese
> characters are absent in lohit-bengali fonts.
>
> I suspect (and therefore, may be wrong here) RH was being politically
> correct in using different styled fonts for same code block and naming
> them lohit-bengali and lohit-assamese.

Comparing the two fonts (from ttf-indic-fonts 1:0.5.12) in Fontforge
with its ”Compare Fonts...” feature gives:
1. Font names differ
2. Two additional glyphs in Lohit-Assamese (uni09E4 and uni09E5)
3. Some minor differences between some glyphs (u09E2, u09F2 and u09F4..u09F8)

The additional glyphs are empty. The minor differences are cosmetic as
far as I can see and nothing script/language specific. The only real
difference I can see in this version of the fonts is actually the
naming...

To me it seems Lohit-Assamese is simply Lohit-Bengali with two
additional “glyphs”. And if I were upstream I would only keep one font
around. (If the reason for the two fonts are political/ideological the
font could be named something completely different without using
certain language names.)

Debian-wise, I would propose to package the two fonts in the same
package (fonts-lohit-beng) for now and wait for upstream to
comment/fix the issue.

Just my two “annas”,
Kess

p. s. Thanks for all your packaging work, the font packages are
starting to get some semblance of order now. ;-)