Myanmar Language Enabling in PangoOSS

1
Myanmar Language Enabling inPango/OSS

• Tin Myo Htet

Seminar on Myanmar Natural Language
Processing May 4, 2006, Conference Hall, Myanmar
Info-Tech
2
GNU/Linux Desktop Structure and
Internationalization
3
Pango

• Pan (All / Greek) Go (Language / Japanese)
• Powerful modular Unicode text rendering engine
• Almost all GTK/Gnome applications use Pango for
  text rendering
• Enabling Myanmar Unicode in Pango gt Enabling
  Myanmar Unicode in GTK/Gnome app.
• Not just for GTK, possible to develop any
  application for Myanmar Unicode.

4
Pango Architecture

• Unicode as common character set
• Modularity
• Rendering system independent
• Low level process of rendering text
  Itemization, Boundary resolution, Shaping, Line
  breaking, Rendering.
• High level PangoLayout handles all above
  processing, and Interactive editing like cursor
  positioning.

5
Enabling Myanmar Unicode in Pango

• Enabling Myanmar Unicode in Pango includes
• Pango Shaper Module
• Myanmar OpenType Font

6
Pango Shaper Module

• Based on Shaper module for Khmer script
• Develop State Table to determine the length of
  well formed Myanmar Syllable.
• Process of shaper module includes getting
  character class, determining syllable boundary,
  reordering, adding dotted circle for wrong
  encoding sequence, assigning correct OpenType
  feature tag according to character class.

7
Syllable?

• Orthographic not Phonological syllable
• Example for the word rEÅav
• Phonologicalref ? w ? av
• Orthographicalr ? ef ? w ? av (or)rEÅ ?
  avThe latter needs more processing for
  reordering but choose this as it can be reused
  for line breaking.

8
Myanmar OpenType Font

• Existing Myanmar OpenType font Myanmar1 cant
  be used as it implement rendering engine specific
  features like reordering.
• Softwares used to create Myanmar OpenType font
  are-FontForge (http//fontforge.sourceforge.net/)
  TrueType/OpenType Tools esp. OTComp,
  TrueTypeViewer (http//home.kabelfoon.nl/slam/fon
  ts/index.html)
• As OpenType specification for Myanmar script has
  not been defined, effort has been made in
  choosing the features to be compatible with
  future support of Uniscribe Engine in MS platform.

9
OpenType features

• Features used in Myanmar OpenType font are-
• abvf ( eg. kinzi vuFm )
• pref ( eg. medial ra Mu )
• blwf ( eg. medial wa uG )
• pstf ( eg. medial ya us )
• abvs ( eg. vowel i anusvara wm0wðom )
• pres ( eg. medial ra wide consonant Mu )
• blws ( eg. medial ha vowel u rI )
• psts ( eg. tall vowel aa cg )
• abvm ( eg. kinzi tall vowel aa ocFg )
• blwm ( eg. subscript ta rEÅav? rdwåL )

10
What supports / Known Issues

• Render correctly Myanmar Unicode conforming
  Unicode 4.x
• Sanskrit extension has not been implemented as
  encoding is not clearly defined yet although
  glyphs are added already.
• Has to improve in line-breaking, cursor
  positioning, etc. specific to Myanmar script as
  now using general one supported by Pango.