Re: [Freebangfont-devel] Re: [Issue N22662] Bengali rendering bugs in Qt 3.2 beta

[qt-bugs]

On Wednesday, 21. May 2003 13:16 Deepayan Sarkar wrote:
>
> Hi,
>
> wow, that was fast ! I'll report on the rest when I finish compiling
> (which may take a while), but let me answer you on the a + ya-phala
> thing.
>
> [Others on the FBF list, please correct me if I have said something
> wrong]
>
> On Wednesday 21 May 2003 05:01, address@hidden wrote:
> > Ok, that's something new for me. I always thought combinations of
> > independent vowels+halant were forbidden, and that's how I handled
> > it
> > in Qt. Looks like I need an exception for bengali.
>
> You are correct, from a linguistic point of view.
>
> The problem was that there was no official way to write english words
> like 'at' (because there was no vowel with the correct sound -- I
> believe Devanagari doesn't have one even now). I guess some smart guy
> decided that Bengali should have the ability to do this, and
> essentially added a new vowel to the language. Unfortunately, no new
> character was created to represent this vowel, instead two completely
> arbitrary combinations of existing symbols were assigned to represent
> this sound, namely what are referred to above as
>
> Vowel_A_zophola_AA and Vowel_E_zophola_AA
>
> Both are completely illegal constructs in classical Bengali.

Ok, so I was mostly correct ;-)
>
> > The faq entry you quote is not 100% clear to me. Does this mean any
> > combination of
> >
> > independent vowel + halant + ya + -aa
> >
> > forms a valid syllable in bengali?
>
> No, as far as I know, no other vowel should have this construct (but
> such combinations would be illegal anyway, so personally I wouldn't
> care how they are rendered).

Usually illegal constructs in indic languages are rendered by prefixing
the illegal character with the dotted circle (0x25cc). Every indic font
should define this glyph. This is how Microsoft handles it and how we
do it.

> > What about the general
> > vowel + halant + consonant + matra
> > case?
>
> Nope, these should be illegal as well. Basically (as far as I know)
> Vowel_A_zophola_AA and Vowel_E_zophola_AA are two very explicit
> exceptions to the otherwise correct general rule you already have.
> (These 2 can be followed by combining marks like candrabindu, bisarga,
> etc, but I don't think that's an issue here.)

Ok, then I can restrict the checking a bit more and only allow the
Halant after Vowel_A and Vowel_E.

> > I've worked around this for now by treating an Independent Vowel at
> > the
> > start of a syllable identical to a consonant for syllable breaking
> > rules in Bengali, so your example renders correctly. I do however
> > not
> > know if this breaks anything else.
>
> It shouldn't. The independent vowel + hasanta construct is illegal
> except for these 2 exceptions, so they should never occur otherwise in
> valid bengali text. I don't know what the unicode/opentype rules are
> when it comes to displaying something invalid, but I don't foresee any
> practical problems.
>
> Deepayan
>
> P.S.: There's another related issue (which I don't think has been
> completely resolved yet), which is how to render combinations of
>
> ra + hasanta + ya (09B0 + 09CD + 09AF)
>
> should it be "reph + ya" or "ra + zophola" (the second is rare, but
> needed again for writing english words like 'rat'). This is an
> ambiguity in the language, and at some point, Unicode should come up
> with a recommended way to represent this. I'll let you know when I
> come to know of anything concrete.

As far as I understand the rules it should be reph+ya. If you add a Zero
Width Joiner (ra + halant + zwj + ya), you should probably get the
second case. But then again I don't speak the language and will
implement whatever people agree upon ;-)

Cheers,
Lars


--
Lars Knoll, Senior Software Engineer
Trolltech AS, Waldemar Thranes gt. 98, N-0175 Oslo, Norway