Pango support for Opentype

[victor_lin21]

Ok, let start... This is not confirm, yet. Inspection undergoing. This is already confirmed. Start from Pango 1.2.* support Opentype Standard Implementation procedure. Refer next post.

http://library.gnome.org/devel/pango/
http://www.pango.org/
http://en.wikipedia.org/wiki/Pango
http://behdad.org/
http://packages.ubuntu.com/hardy/libs/pango-graphite

Quick Test Run
===========

Wrong Input Sequence

ေန
ြမ
--ု + --ိ


Right Input Sequence

နေ
မြ
--ိ + --ု

[victor_lin21]

This is to update that start from Pango 1.20, it support Standard Opentype Specification font registry lookup feature tag. Especially, working of GSUB layout operation;

- liga Standard ligature substitution
- clig Contextual ligature substitution

The following screenshots are testing on all current U51 Opentype font on Pango 1.20.* where now shipped version of Fedora 9 and Ubuntu 8.08.

Pango Only
=============
First, I make sure that, my system has only Pango. That mean, this stock shipment of GNOME desktop where no pango-graphite is not being patched/installed.

pango_only_fedora.png

pango_only_ub.png

One-Font-At-A-Time
====================
I make sure that, each test cases, my system only has one Burmese U51 font at a time.

Myanmar2 and Myanmar3 Testing

mm2_U51.png

mm3_U51.png

Padauk (V2.4) Testing

padauk_U51.png

Parabaik Testing

parabaik_U51.png

MyMyanmar Testing

mymyanmar_U51.png

Bug Reports
Parabaik is not working the ligature substitution because of the usage of require ligature "rlig" tag. Either of "clig" or "liga" shall fix the bug on Pango's Opentype Standard Implementation.

Myanmar2 is not working the ligature substitution because of the usage of required ligature "rlig" tag. Either of "clig" or "liga" shall fix the bug on Pango's Opentype Standard Implementation.

Supporting Evident

pango_opentype_ligature.png

Summary
============
From now on, GNU/Linux or any Unix alike variant system which make use of GNOME desktop (underlying Pango) shall treat the same as Microsoft Window XP (SP2 or VOLT usp10.dll patched) and Vista. Which mean no additional requirement for default system in order to display Unicode 5.1 formatted text. The input method for GNU/Linux shall investigate in further but very much and quick solutions possibility out there.

Reference
=============
http://www.sankholin.com/pango-opentype/
http://www.unimm.org/cms/node/23

Credits
==========
The first time report to me is who AK(47) Htet on Sat, Jul 5, 2008 at 5:30 PM, where he unintentionally installed just Parabaik/Padauk on his Ubuntu box and happened to visit my.wikipedia.org, and said to me that:

hey, dude..
i can see the stuffs

By then, I follow up by updating all my boxes to date and made investigation.

EoF

[maungwiki]

I tried to test with Myanmar 3 and it displays quite well except some incorrect rendering with Nga That Aut ka Myint. for example And=Hnit (Na Nge hla Htoo aut ka myint). Mathematics=Thinchar (Thin Char)

Could you explain me what wrong with Myanmar 3?

[victor_lin21]

I tried to test with Myanmar 3 and it displays quite well except some incorrect rendering with Nga That Aut ka Myint. for example And="Hnit" (Na Nge hla Htoo aut ka myint). Mathematics="Thinchar" (Thin Char)
Could you explain me what wrong with Myanmar 3?

Pls take note that, for Myanmar 3 font and input type suite, you just have to type as in syllable structure and mean.

Hnit = na ngal + ha htoe + nga + a thet + aut ka myint
Thinchar = ta + nga + a thet + kha + ya pint + yay char

Pls copy/paste your text here in reply for further analysis, if the problem still persist.

[maungwiki]

Sure, I type in syllable order. You will also see such kind of error on above discussion screenshot mm3_U51.png
Like Shay Nay U Phar Hnint Daw Su

[victor_lin21]

Sure, I type in syllable order. You will also see such kind of error on above discussion screenshot mm3_U51.png
Like Shay Nay U Phar Hnint Daw Su

1) The article is stating that it has been converted from some format to Unicode 5.1.
2) Thus, the error assume came from what previous original article writer's typo/syllable entry error.
3) Do note and read The need of dotted-circle and spell checker in concept article for details.

နှင့် သင်္ချာ

ရှေ့နေဦးဖာနှင့် မိခင်မှာ ဒေါ်ခင်စု

correct sequence => နှင့် => \u1014\u103E\u1004\u103A\u1037
incorrect sequence => နှင့် => \u1014\u103E\u1004\u1037\u103A (this is current situation in Wiki entry)

The syllable entry should be, --် a thet (u103A) follow by --့ aut ka myint (u1037). Not in reverse order. If typo in reverse so, Myanmar3 internal font dotted circle prevented for display. In principal, dotted circle check in this case should not enforce at font level. But by doing so as long as in right manner and not to interfere with Unicode 5.1 and UTN11-2 rule then it said so to be ok. That's, you may assume what Myanmar3 extra feature for syllable support inside font. At most extreme situation, this may assume as typographic feature or font designer specific feature.

hnint.jpg

So, in contrary, if you happen to test with Padauk 2.4 for above text, it will show no error by font display. But there is an typo error as you have been already experienced in above.

hint_padauk.jpg

In summary, syllable checking inside font or syllable checker should be standalone application. But so far, there is no such Burmese standalone syllable/spelling checking software yet published in public than font itself internally checking. The main principal of dotted circle prevention and reordering is major aspectof the Vowel E and the different forms of Medial case by implmenting at rendering engine but again, the font is doing this. That's what our current happening of Myanmar Burmese Opentype font and situation. Currently, you may refer : http://unimm.blogspot.com/2007/12/reord ... anmar.html article and, within few days later I will write "The Present Opentype Reordering for Burmese" new article. So, still tune.

[friendklay]

Sure, I type in syllable order. You will also see such kind of error on above discussion screenshot mm3_U51.png
Like Shay Nay U Phar Hnint Daw Su

1) The article is stating that it has been converted from some format to Unicode 5.1.
2) Thus, the error assume came from what previous original article writer's typo/syllable entry error.
3) Do note and read The need of dotted-circle and spell checker in concept article for details.

နှင့် သင်္ချာ

ရှေ့နေဦးဖာနှင့် မိခင်မှာ ဒေါ်ခင်စု

correct sequence => နှင့် => \u1014\u103E\u1004\u103A\u1037
incorrect sequence => နှင့် => \u1014\u103E\u1004\u1037\u103A (this is current situation in Wiki entry)

The syllable entry should be, --် a thet (u103A) follow by --့ aut ka myint (u1037). Not in reverse order. If typo in reverse so, Myanmar3 internal font dotted circle prevented for display. In principal, dotted circle check in this case should not enforce at font level. But by doing so as long as in right manner and not to interfere with Unicode 5.1 and UTN11-2 rule then it said so to be ok. That's, you may assume what Myanmar3 extra feature for syllable support inside font. At most extreme situation, this may assume as typographic feature or font designer specific feature.

hnint.jpg

So, in contrary, if you happen to test with Padauk 2.4 for above text, it will show no error by font display. But there is an typo error as you have been already experienced in above.

hint_padauk.jpg

In summary, syllable checking inside font or syllable checker should be standalone application. But so far, there is no such Burmese standalone syllable/spelling checking software yet published in public than font itself internally checking. The main principal of dotted circle prevention and reordering is major aspectof the Vowel E and the different forms of Medial case by implmenting at rendering engine but again, the font is doing this. That's what our current happening of Myanmar Burmese Opentype font and situation. Currently, you may refer : http://unimm.blogspot.com/2007/12/reord ... anmar.html article and, within few days later I will write "The Present Opentype Reordering for Burmese" new article. So, still tune.

Hello Victor and Muangwiki,
there is a bug in MediaWiki. MediaWiki normalization all Unicode Data so that we can search for syllable clusters even if we typed a different sequence. However the way it normlizes this seems to be different from UTN-11. It is however following the UCD. The normalization data in UCD is not in sync with UTN-11. There is also an issue with Kinzi in MediaWiki. Pls check Bug 2399, starting from comment #35 is about Myanmar. I can understand why MediaWiki normalizes Athet and AuKaMyint the way it does, infact I prefer that sequence better. But I think it is plain wrong for Kinzi. Trying entering သင်္ချာ and see what happens in MediaWiki.

Victor, could you follow up on that 'bug' in MediaWiki. I am finding it hard to shift through the UCD to find a sequence for Kinzi.
Ref:
http://www.unicode.org/faq/normalization.html#8
http://www.unicode.org/Public/5.1.0/ucd/
Many Thanks,
Ravi.

[maungwiki]

Hi Ravi,
Thank you very much for your informative explanation.

[victor_lin21]

For UTN11-2 and Normalization issue (Mediawiki), please divert to trail the following topic.
viewtopic.php?f=83&t=1046