Re: Malayalam rendering

Re: Malayalam rendering

[Kenichi Handa]

In article <874r6ukjqy.fsf@vsnl.net>, gnu@vsnl.net (Ramakrishnan M) writes:
> I found some minor problems in emacs rendering of afew
> compound characters. I have a patch which fixes the rendering.
> They are minor changes (2 lines).

Thank you.  I installed that change.

> I guess since these are small changes, I need not go through
> any paper work for any copyright assignment to FSF ??

Right.

> I am attaching the words Malayalam, and equivalent of "Hello"
> in Malayalam. It is a UTF-8 file. Thanks

Thank you.  I added that in the HELLO file.

> I have put afew screenshots at http://www.hackGNU.org/emacs-mal/

This screenshot contains many empty boxes.
    http://www.hackGNU.org/emacs-mal/Screenshot-1.png
What are they?  Is it a bug that is fixed by the above
change?

---
Ken'ichi HANDA
handa@m17n.org

Re: Malayalam rendering

[Ramakrishnan M]

 || On Mon, 24 Feb 2003 09:53:06 +0900 (JST)
 || Kenichi Handa <handa@m17n.org> wrote: 

 kh> Thank you.  I installed that change.

Thanks!

 >> I have put afew screenshots at http://www.hackGNU.org/emacs-mal/

 kh> This screenshot contains many empty boxes.
 kh>     http://www.hackGNU.org/emacs-mal/Screenshot-1.png
 kh> What are they?  Is it a bug that is fixed by the above
 kh> change?

The above patch fixes some of the problems in that screenshot. But that text
as shown in that screenshot was not written using Emacs. It was generated using
some other tools, I guess by someone else (It is the translation of GNU GPL in 
Malayalam). There are problems in the generated file. So I guess we don't have
to worry. I will try to write the same text natively in Emacs.

best regards
-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!

Re: Malayalam rendering

[Kenichi Handa]

In article <87isva4bmn.fsf@vsnl.net>, gnu@vsnl.net (Ramakrishnan M) writes:
kh>  This screenshot contains many empty boxes.
kh>      http://www.hackGNU.org/emacs-mal/Screenshot-1.png
kh>  What are they?  Is it a bug that is fixed by the above
kh>  change?

> The above patch fixes some of the problems in that
> screenshot. But that text as shown in that screenshot was
> not written using Emacs. It was generated using some other
> tools, I guess by someone else (It is the translation of
> GNU GPL in Malayalam). There are problems in the generated
> file. So I guess we don't have to worry. I will try to
> write the same text natively in Emacs.

As I found this file on the net:
   http://fsf.org.in/projects/malayalam-gpl/gpl-ml-unicode-pre.txt
I checked the contents.  Those empty boxes are ZWJ
characters.  But, it seems strange that we need so many ZWJ
for Malayalam writing.

---
Ken'ichi HANDA
handa@m17n.org

Re: Malayalam rendering

[Ramakrishnan M]

 || On Mon, 24 Feb 2003 14:45:24 +0900 (JST)
 || Kenichi Handa <handa@m17n.org> wrote: 

 kh>    http://fsf.org.in/projects/malayalam-gpl/gpl-ml-unicode-pre.txt
 kh> I checked the contents.  Those empty boxes are ZWJ
 kh> characters.  But, it seems strange that we need so many ZWJ
 kh> for Malayalam writing.

One issue is that the Malayalam section as it is, now in the Unicode
standard has problems. There is a set of 5 characters called "Chillaksharam"
which are not there in the standard. TDIL (some organization in Govt of India
which does language related work) has proposed some new changes, which 
sort of, solves this issue. So current tools handle "chillaksharams" is their
own non-standard way. One way is <base char>+<halant>+<ZWJ>. This might be a 
reason you see alot of ZWJ in that text. I may be wrong though...

-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!

Re: Malayalam rendering

[Kenichi Handa]

In article <87smuecj1w.fsf@vsnl.net>, gnu@vsnl.net (Ramakrishnan M) writes:
> One issue is that the Malayalam section as it is, now in the Unicode
> standard has problems. There is a set of 5 characters called "Chillaksharam"
> which are not there in the standard. TDIL (some organization in Govt of India
> which does language related work) has proposed some new changes, which 
> sort of, solves this issue. So current tools handle "chillaksharams" is their
> own non-standard way. One way is <base char>+<halant>+<ZWJ>. This might be a 
> reason you see alot of ZWJ in that text. I may be wrong though...

I see.  Apart from the correctness of that non-standard way,
if the Unicode standard requires that such a sequence should
be rendered correctly, Emacs should support it.  Do you know
what Unicode requires?  Perhaps, supporting ZWNJ is
necessary as well.

Kawabata-san, is it difficult to support ZWJ and ZWNJ?

---
Ken'ichi HANDA
handa@m17n.org

Re: Malayalam rendering

[Ramakrishnan M]

 || On Mon, 24 Feb 2003 16:54:45 +0900 (JST)
 || Kenichi Handa <handa@m17n.org> wrote: 

 kh> I see.  Apart from the correctness of that non-standard way,
 kh> if the Unicode standard requires that such a sequence should
 kh> be rendered correctly, Emacs should support it.  Do you know
 kh> what Unicode requires?  Perhaps, supporting ZWNJ is
 kh> necessary as well.

Let me go through the unicode docs and get back to you.

-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!

Malayalam -- more fixes

[Ramakrishnan M]

[-- Attachment #1: Type: text/plain, Size: 99 bytes --]


Attached is a patch which fixes rendering of consonent + vowel u.
Again they are minor changes..


[-- Attachment #2: patch-fixes --]
[-- Type: application/octet-stream, Size: 1480 bytes --]

[-- Attachment #3: Type: text/plain, Size: 123 bytes --]



-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!

[-- Attachment #4: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Malayalam rendering

[Ramakrishnan M]

 || On Mon, 24 Feb 2003 16:54:45 +0900 (JST)
 || Kenichi Handa <handa@m17n.org> wrote: 

 kh> I see.  Apart from the correctness of that non-standard way,
 kh> if the Unicode standard requires that such a sequence should
 kh> be rendered correctly, Emacs should support it.  Do you know
 kh> what Unicode requires?  Perhaps, supporting ZWNJ is
 kh> necessary as well.


I did some web search on this topic and found[1] that, those ZWJs
were artifacts of convertions from ISCII to Unicode. 

It seem to me that consonant + halant + ZWJ == half consonent
(valid for those 5 half consonants (chillaksharam)) should be
supported. May be emacs-unicode hackers have something to say on
this.

[1] http://www.unicode.org/faq/indic.html#16
-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!

Re: Malayalam -- more fixes

[Kenichi Handa]

In article <87vfzaoxir.fsf_-_@vsnl.net>, gnu@vsnl.net (Ramakrishnan M) writes:
> Attached is a patch which fixes rendering of consonent + vowel u.
> Again they are minor changes..

Installed, thank you.

> I did some web search on this topic and found[1] that, those ZWJs
> were artifacts of convertions from ISCII to Unicode. 

> It seem to me that consonant + halant + ZWJ == half consonent
> (valid for those 5 half consonants (chillaksharam)) should be
> supported. May be emacs-unicode hackers have something to say on
> this.

> [1] http://www.unicode.org/faq/indic.html#16

Kawabata-san, what do you think?

---
Ken'ichi HANDA
handa@m17n.org

Re: Malayalam -- more fixes

[Ramakrishnan M]

[-- Attachment #1: Type: text/plain, Size: 512 bytes --]

 || On Mon, 24 Feb 2003 20:22:36 +0900 (JST)
 || Kenichi Handa <handa@m17n.org> wrote: 

 kh> In article <87vfzaoxir.fsf_-_@vsnl.net>, gnu@vsnl.net (Ramakrishnan M) writes:
 >> Attached is a patch which fixes rendering of consonent + vowel u.
 >> Again they are minor changes..

 kh> Installed, thank you.

Attached is another patch. Sorry.. I should have sent one patch with
all changes.

-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!


[-- Attachment #2: patch-new --]
[-- Type: application/octet-stream, Size: 1392 bytes --]

[-- Attachment #3: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Malayalam -- more fixes

[Ramakrishnan M]

 || On Mon, 24 Feb 2003 20:22:36 +0900 (JST)
 || Kenichi Handa <handa@m17n.org> wrote: 

 >> [1] http://www.unicode.org/faq/indic.html#16

 kh> Kawabata-san, what do you think?

BTW, Chapter 9 of the Unicode book, doesn't say about this
in the context of malayalam. But in the case of Devanagari,
it says that half-consonents "may be" encoded as 

 <consonent>+<halant>+<ZWJ> 

(Please see Table 9-1 in that chapter at www.unicode.org)

Does the current Devanagari implementation take care of this?

-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!

Re: Malayalam -- more fixes

[Ramakrishnan M]

[-- Attachment #1: Type: text/plain, Size: 525 bytes --]

 || On 24 Feb 2003 18:57:53 +0530
 || gnu@vsnl.net (Ramakrishnan M) wrote: 

 rm> BTW, Chapter 9 of the Unicode book, doesn't say about this
 rm> in the context of malayalam. But in the case of Devanagari,
 rm> it says that half-consonents "may be" encoded as 

 rm>  <consonent>+<halant>+<ZWJ> 

One way to handle "chillaksharam" (or half-consonants) is to input
the following.

<consonent>+<vowel U>+<halant>

The GNOME guys who do Malayalam support is going this way. I added
the support for this. Please find the patch.


[-- Attachment #2: patch-new --]
[-- Type: application/octet-stream, Size: 2517 bytes --]

[-- Attachment #3: Type: text/plain, Size: 122 bytes --]


-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!

[-- Attachment #4: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: Malayalam rendering

[川幡 太一]

Hello, 

When I was back to home and read the mails, I was surprised to see so
many mail exchanged...

>>>>> In <87r89yox1u.fsf@vsnl.net> (Re: Malayalam rendering), 
>>>>> Mr. Ramakrishnan wrote

> I did some web search on this topic and found[1] that, those ZWJs
> were artifacts of convertions from ISCII to Unicode. 

> It seem to me that consonant + halant + ZWJ == half consonent
> (valid for those 5 half consonants (chillaksharam)) should be
> supported. May be emacs-unicode hackers have something to say on
> this.

It seems to me, as I've seen from the Table 9-1 of Unicode standards,
consonant + halant + ZWJ is used to form the half form of the
consonants.  Is it common to use these characters in Malayalam?  I
thought that the glyphs shown in Table 9-1 is only used in language
textbooks and not at the common usage.  

Anyway, I thinkg it is easy to support them if the appropriate
corresponding glyphs are already prepared in indian-glyph.  I think it
is just add the new entries for them in mlm-char-glyph variable, isn't
it?

-- 
---------------------------------------
 KAWABATA, Taichi (kawabata@m17n.org)
---------------------------------------

Re: Malayalam -- more fixes

[川幡 太一]

Hello, 

>>>>> In <877kbp95u3.fsf@vsnl.net> (Re: Malayalam -- more fixes), 
>>>>> Ramakrishnan wrote:

> Attached is another patch. Sorry.. I should have sent one patch with
> all changes.

I put the comment of `;; ?' at the place where I'm not sure about the
ligature rule.  If you check them and seem they are all right, would
you please remove these "unsure" comments?  I really appreciate it...

-- 
---------------------------------------
 KAWABATA, Taichi (kawabata@m17n.org)
---------------------------------------

Re: Malayalam -- more fixes

[Ramakrishnan M]

 || On Tue, 25 Feb 2003 01:37:20 +0900
 || 川幡 太一 <kawabata@m17n.org>( Taichi KAWABATA ) wrote: 

 川太> I put the comment of `;; ?' at the place where I'm not sure about the
 川太> ligature rule.  If you check them and seem they are all right, would
 川太> you please remove these "unsure" comments?  I really appreciate it...

I have changed it. Thanks.

-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!

Re: Malayalam rendering

[Ramakrishnan M]

 || On Tue, 25 Feb 2003 01:33:35 +0900
 || 川幡 太一 <kawabata@m17n.org>( Taichi KAWABATA ) wrote: 

 川太> Hello, 
 川太> When I was back to home and read the mails, I was surprised to see so
 川太> many mail exchanged...

Yeah.. I am on vacation for a week and got some time to work on
interesting stuff only now. So I thought this is the best time to
see Malayalam implementation on Emacs.

 川太> It seems to me, as I've seen from the Table 9-1 of Unicode standards,
 川太> consonant + halant + ZWJ is used to form the half form of the
 川太> consonants.  Is it common to use these characters in Malayalam?  I
 川太> thought that the glyphs shown in Table 9-1 is only used in language
 川太> textbooks and not at the common usage.  

Yes, the five half-consonents are very common in Malayalam. Without that
the meaning of some words just change. i.e <consonent>+<halant> has
different meaning compared to <half-consonant> for those 5 characters.

 川太> Anyway, I thinkg it is easy to support them if the appropriate
 川太> corresponding glyphs are already prepared in indian-glyph.  I think it
 川太> is just add the new entries for them in mlm-char-glyph variable, isn't
 川太> it?

Yes. I think so.

-- 
    Ramakrishnan M              (http://www.hackGNU.org/)
    Use Free Software -- Help stamp out Software Hoarding!