Complex Script Functionality: Difference between revisions

From Scribus Wiki
Jump to navigation Jump to search
Line 238: Line 238:


===Language Script & Language System tags:===
===Language Script & Language System tags:===
* Tibetan:tibt:dflt (default)
* Tibetan:tibt:TIB
* Tibetan:tibt:TIB
* Dzongkha:tibt:DZN
* Dzongkha:tibt:DZN
* Sanskrit:tibt:SAN
* Ladakhi:tibt:LDK
* Ladakhi:tibt:LDK
* Tshangla:tibt:TSJ 
* Khengkha:tibt:XKF   
* Bumthangkha:tibt:KJZ
* Balti:tibt:BFT


===OpenType Features===
===OpenType Features===

Revision as of 09:16, 20 December 2012

This page relates to Bug 0001547: - Support for Indic Scripts and Metabug 0003965: Support for non-latin languages].

Things we require for testing Complex Script Functionality

  • Sample text files encoded in UTF8
  • Sample screenshots of Scribus 1.3.12, 1.3.4 or 1.3.5cvs rendering these incorrectly.
  • Sample fonts that we can get that are freely available that have the ability to be used for DTP.
  • Sample screenshots of any application rendering the same text correctly

An important point to note is we need comparable information. Same texts, same fonts etc and screenshots (png or tiff please) of said texts and fonts. --Cbradney 18:40, 25 Feb 2005 (UTC)

  • Other usefull contribution would be mapping of the languages you're interested in with its script/lang pair in OpenType Font style.

example : French:latn:FRA

To help, you can have a look at http://www.microsoft.com/typography/developers/OpenType/scripttags.aspx for script tags and http://www.microsoft.com/typography/developers/OpenType/languagetags.aspx for lang tags. --Pmarchand 13:46, 8 April 2007 (CEST)

Please try to only list the best fonts for each script for use in printed publications. (Not screen/web fonts).

I don't know anything about quality, but Fedora 9 has many of these fonts available in their repository (meaning they must be free): arabic, bengali, gujarati, hindi, malayalam, oriya, punjabi, sinhala, tamil, and telugu. – Greg P

Arabic & Urdu

Font: Free fonts for Arabic / Urdu are available by Crulp and by PakType. Licensing is GPL at PakType and Crulp has also very loose license.

Bengali

Font:


SolaimanLipi is a free Bengali font available in sourceforge

Language, Script & Language System tags:

  • Assamese:beng:ASM
  • Bengali:beng:BEN

Open Type features used for Bengali script


Screenshots:


Please expand this section

Devanagari

The Devanagari script is used for writing many languages in India and Nepal including Sanskrit, Hindi and Nepali.

Fonts:

  • Chandas - Devanagari Unicode Open Type font with 4347 glyphs: 325 half-forms, 960 half-forms context-variations, 2743 ligature-signs. It is designed especially for Vedic and Classical Sanskrit but can also be used for Hindi, Nepali and other modern Indian languages. The font includes Vedic accents and many additional signs and so provides maximal support for Devanagari script. GPL License.
  • Madan - Open Type font with Nepali glyphsets.Developed by Madan Puraskar Pustakalaya and released under GPL.

Language, Script & Language System tags:

  • Bhojpuri:deva:BHO
  • Hindi:deva:HIN
  • Marathi:deva:MAR
  • Nepali:deva:NEP
  • Sanskrit:deva:SAN

OpenType features used for Devanagari script

Screenshots:

Current Rendering by Scribus

Desired Rendering

Gujarati

  • Please expand this section

Fonts:


Language, Script & Language System tags:


Screenshots:

Kannada

  • Kannada is used in south India

Fonts:
Lohit Kannada https://fedorahosted.org/lohit (GPL Font)


Language, Script & Language System tags:
Kannada KAN


Screenshots:
Scribus rendering http://wiki.scribus.net/index.php/Image:Scribus-kannada.png Desired Rendering http://wiki.scribus.net/index.php/Image:Kwrite-kannada.png

Khmer

  • Khmer is used in Cambodia.

Fonts:


Language, Script & Language System tags:

  • Khmer KHM

Screenshots:
Scribus rendering (incorrect): Khmer-scribus.png

gEdit rendering (correct): Khmer-gedit.png

Lao

  • Please expand this section

Fonts:


Language, Script & Language System tags:


Screenshots:

Malayalam

Fonts:
Lohit Malayalam https://fedorahosted.org/lohit , AnjaliOldLipi http://varamozhi.sourceforge.net/fonts/AnjaliOldLipi-0.710.ttf (GPL fonts)


Language, Script & Language System tags:
Malayalam MAL


Screenshots:
Scribus rendering scribus_ml.png Desired Rendering des_re_ml.png

Myanmar

Please expand this section

Font:


Language, Script & Language System tags:

  • Burmese:mymr:BRM

OpenType Features used for Myanmar script:



Screenshots:

Oriya

Font:

Language, Script & Language System tags:

  • Oriya:orya:ori

Open Type features used for Oriya script


Screenshots: Scribus rendering Scribus or1.png Desired Rendering Scribus or2.png UTF8: <utf8>ଜ୍ଞାନକୋଷ ଭାଷାରେ </utf8>

Sinhala

Fonts:

  • Unicode Range: 0D80 - 0DFF SInhala Unicode Range
  • Fonts freely available in both Linux and MS Windows from here
  • Typing order: Left ==> Right

Language, Script & Language System tags:


Screenshots:

  • Sinhala Unicode Issues for Scribus Version 1.3.3.12
    • Issues in MS Windows

1. Not Supporting for Sinhala Charters

Example Charters
 අ 0D85 , ෆ 0DC6 ......
    • Issues in Linux
  • Sinhala Unicode Issues for Scribus Version 1.3.6SVN
    • Issues in MS Windows
Scribus rendering http://wiki.scribus.net/index.php/Image:Scribus_1.3.6svn(ms_windows)-sinhala.png
Desired Rendering http://wiki.scribus.net/index.php/Image:Notepad-Sinhala.png
    • Issues in Linux

Tamil

  • Please expand this section

Fonts:


Language, Script & Language System tags:


Screenshots:

Telugu

  • Please expand this section

Fonts:


Language, Script & Language System tags:


Screenshots:

Thai

  • Please expand this section

Fonts:
Seems to work fine on a variety of fonts (compiled for OSX), including the following:

  • JS-Jukaphan
  • PSLxText
  • PSL TextSP
  • Thonburi (OSX default font for Thai)
  • TH Krub
  • TH K2D July8


Note: This does not work with Tahoma (the tone marks and vowels overlay each other)

Language, Script & Language System tags:


Screenshots:

Tibetan & Dzongkha

This relates to Bug 0004452: No support for Tibetan Script - sample UTF-8 test file and PDF file attached to that bug.


Tibetan Script:

Fonts:

  • Jomolhari, a free Tibetan script font suitable for publishing is available from the Free Tibetan Fonts project - OFL License. The font was originally designed for publishing traditional Buddhist texts in Classical Tibetan (chos skad) - but can also be used for modern Tibetan, Dzongkha and Ladakhi text. The font works well in OO.org 2.4.x & Inkscape.

Language Script & Language System tags:

  • Tibetan:tibt:dflt (default)
  • Tibetan:tibt:TIB
  • Dzongkha:tibt:DZN
  • Sanskrit:tibt:SAN
  • Ladakhi:tibt:LDK
  • Tshangla:tibt:TSJ
  • Khengkha:tibt:XKF
  • Bumthangkha:tibt:KJZ
  • Balti:tibt:BFT

OpenType Features

The following OpenType features should be processed to support Tibetan script:

  • ccmp - Glyph Composition / Decomposition (Composition = GSUB lookup type 4 & Decomposition = GSUB lookup type 2).
  • blws - Below-base Substitutions (GSUB lookup type 4)
  • abvs - Above-base Substitutions (GSUB lookup type 4)
  • calt - Contextual Alternates (GSUB lookup type 6)
  • blwm - Below-base Mark Positioning (GPOS lookup type 4,5)
  • abvm - Above-base Mark Positioning (GPOS lookup type 4,5)
  • kern - Kerning - (GPOS lookup type 2 or 8)

Note: The current version of the Jomolhari font uses only ccmp, blws, abvs, calt, and kern features. In kern only GPOS lookup type 2 (pair-adjustment) is used in this font ~ so this is the only kind of GPOS lookup used.


Tibetan Script Rendering Examples:

Test of ccmp feature used for Decomposition (GSUB lookup type 2):

The string U+0F43 U+0F77 (གྷཷ) should render like the first example:

Correct rendering

However, in Scribus, U+0F43 U+0F77 (གྷཷ) currently renders like this:

rendering in Scribus

In this case, the ccmp feature should decompose U+0F77 into seperate glyph components to go below and above the glyph for the base character (U+0F43)

Test of ccmp feature used for Composition (GSUB lookup type 4):

The string U+0F42 U+0FB7 U+0F0B U+0F4C U+0FB7 U+0F0B U+0F51 U+0FB7 U+0F0B U+0F56 U+0FB7 U+0F0B U+0F5B U+0FB7 U+0F0B U+0F40 U+0FB5 (གྷ་ཌྷ་དྷ་བྷ་ཛྷ་ཀྵ) should render as follows:

2a. Correct Rendering in MSWord 2003

However, in Scribus, that string currently renders like this:

2b.Rendering in Scribus

In this case if ccmp feature is being applied ligatures should be properly composed.

Furthermore, the strings U+0F42 U+0FB7 U+0F0B U+0F4C U+0FB7 U+0F0B U+0F51 U+0FB7 U+0F0B U+0F56 U+0FB7 U+0F0B U+0F5B U+0FB7 U+0F0B U+0F40 U+0FB5 (གྷ་ཌྷ་དྷ་བྷ་ཛྷ་ཀྵ) and U+0F43 U+0F0B U+0F4D U+0F0B U+0F52 U+0F0B U+0F57 U+0F0B U+0F5C U+0F0B U+0F69 (གྷ་ཌྷ་དྷ་བྷ་ཛྷ་ཀྵ) should render identically as they are canonically equivalent. (U+0F42 U+0FB7 = U+0F43; U+0F4C U+0FB7 = U+0F4D; U+0F51 U+0FB7 = U+0F52; U+0F5B U+0FB7 = U+0F5C; and U+0F40 U+0FB5 = U+0F69.)

Test of blws feature (GSUB lookup type 4)

3a. Correct Rendering in OO.org

Test string consists of the folowing Unicode characters:
U+0F62 U+0F0B U+0F62 U+0F90 U+0F0B U+0F62 U+0F90 U+0FB1 U+0F0B U+0F62 U+0F90 U+0FB1 U+0F74 U+0F62 U+0F92 U+0F0B U+0F62 U+0F92 U+0FB1 U+0F0B U+0F62 U+0F92 U+0FB1 U+0F74 U+0F0D (ར་རྐ་རྐྱ་རྐྱུརྒ་རྒྱ་རྒྱུ།) and

U+0F66 U+0F0B U+0F66 U+0FA6 U+0F0B U+0F66 U+0FA6 U+0FB2 U+0F0B U+0F66 U+0FA6 U+0FB2 U+0F74 U+0F0B U+0F66 U+0FA8 U+0F0B U+0F66 U+0FA8 U+0FB2 U+0F0B U+0F66 U+0FA8 U+0FB2 U+0F74 U+0F0D (ས་སྦ་སྦྲ་སྦྲུ་སྨ་སྨྲ་སྨྲུ།)
3b. Rendering in Scribus

Test of abvs feature (GSUB lookup type 4)

Hebrew with vowels and cantillation

UTF8 Text

בְּרֵאשִׁ֖ית בָּרָ֣א אֱלֹהִ֑ים אֵ֥ת הַשָּׁמַ֖יִם וְאֵ֥ת הָאָֽרֶץ׃

Scribus Screenshot

Fonts

Screenshot of BibleTime