Chris Pratley

1
and
Unicode

• Chris Pratley
• Group Program Manager
• Microsoft Word

2
Overview

• Office Unicode history and strategy
• Implementation
• Benefits of Unicode to Office users
• Demo of Word

3
Office97 Unicode Strategy

• Office97 driving factors
• Customers operate world-wide (US only 40)
• Need to handle multiple code pages in Europe
• Office97 goals
• Enable loss-less file exchange world-wide
• Solve code page problems in Europe
• Development efficiency for Asian and Euro
  versions
• Unified source code base but still different
  executables
• Unified development process
• Delta between language versions shrinks from 18
  to 2 months
• Lay foundation for future

4
Office2000 Unicode Strategy

• Office2000 goals
• Reduce Total Cost of Ownership for large
  corporations
• Single version to deploy and administer globally
• Configurable interface to handle local needs
• Language of User Interface can be changed
• Additional language features can be enabled as
  needed
• Emulate any localized version
• Français, ???, ??, ?????,????, etc.
• Streamline development process further
• Core US team ships global product
• Integrate bi-directional version team (Arabic,
  Hebrew)
• Focus on needs of bilingual and multilingual users

5
Office 10 Unicode Strategy

• Office10 goals
• Finish the globalization work begun in Office2000
• Extend functionality to all applications
• Integrate Complex Scripts support (Indic, Thai,
  Vietnamese)
• ??????, ???????, ???????, Viê?t 
• Streamline development process further
• Single build process from start to finish
• Integrate complex scripts team
• Deepen Unicode support
• Unicode 3.0 languages (??????,  ????, etc.)
• UTF-16 (esp. plane 2 ????????)
• More complex script and limited combining
  diacritic coverage

6
The Word Family Tree
Version
7
Implementation

• Core applications are Unicode internally
• Word, Excel, PowerPoint (Office97)
• Access, Publisher (new in Office2000)
• Databases and drivers are Unicode
• Outlook, FrontPage (new in Office10)
• New Outlook local storage is Unicode

8
Implementation

• Difficulties encountered with Unicode
• Lack of full system support in Win9x
• Every app needed different solution
• MFC-based apps were hardest
• Missing system services (e.g. font-linking)
• Interoperation with code-page based systems
• Educating test team about Unicode
• Testing issues different vs. MBCS
• Lack of expertise in uncommon languages

9
Implementation

• Office shared code services
• Central Win32 Unicode text API wrappers
• Simulate nearly full support on Win9x
• ExtTextOutW and others
• Provide optional font-linked output
• Hardcode preferred fonts by script, style
• User-specified font-fallbacks via reg key (if
  any)
• Font categorization by script range (use
  MLANG.DLL)
• Font substituted if glyph not available
• Word modifies font settings in the document
• Other apps do only at display time
• Insert Symbol dialog (Unicode 3.0 support)

10
Office Users Benefit

• Single binary world-wide
• Shared world-wide file formats
• Multilingual word/data processing
• Unicode HTML
• Unicode e-mail (HTML, RTF, plain)

11
Single Binary

• Easier to deploy, administer
• One set-up image to install world-wide
• One set of service packs for all machines
• All features available in all versions
• Still have local version packages
• Multilingual users can use foreign features
• User Interface language is configurable
• Your language follows you when you travel
• Major cost savings for customers
• Less testing of corporate solutions
• Lower internal tech support costs

12
Single File Format

• Multinational corporations use Office
• Need to exchange documents company-wide
• Office unified file formats via Unicode
• Word95 had 7 different file formats
• Word97 had 1 file format but no editing, layout
  for languages covered by other versions
• Word2000 adds editing, layout, and full-roundtrip
• Word10 adds full complex script support

13
Multilingual Usage

• English Office10 input/display/edit/layout of
• European languages
• any similar left-right scripts if fonts/NLS
  available
• E.g. Canadian Syllabics (Inuktitut), Ethiopic,
  Cherokee
• Some combining diacritic support (African
  languages)
• East Asian languages (including UTF-16
  surrogates)
• Chinese (Traditional and Simplified), Japanese,
  Korean
• Complex Script and Bi-directional scripts (need
  enabled system)
• Arabic (incl. Farsi, Urdu), Hebrew
• Thai
• Hindi, Tamil, Oriya, Telugu, Punjabi, Bengali,
  Gujarati, etc.

14
Multilingual Usage

• Most documents are monolingual
• Most users are bilingual
• Local language
• English
• Optimize UI for using one, two or three languages
• Over 100 supported rare usage
• Detect 20 languages while typing (Word)
• Automatically install and use the correct
  proofing tools
• Plain text I/O in any encoding (Word, Excel)

15
Multilingual Word Processing

• Proofing tool interfaces are Unicode
• SDKs available for 3rd party development
• Tools for over 35 languages available
• European languages, Japanese, Chinese, Korean,
  Arabic, Hebrew, Thai, Hindi
• Spelling, Grammar, Hyphenation, Thesaurus
• Traditional/Simplified Chinese conversion
• Japanese character usage consistency checker
• Hangul/Hanja conversion
• Translation dictionaries (available offline)
• Automatic translation web services

16
Multilingual Data Processing

• Access databases are Unicode
• Hook up to SQL7.x/2000 Unicode databases
• Excel workbooks are Unicode
• Hook up to Unicode databases using OLE-DB
• Create Pivot lists and manipulate Unicode data
• PowerPoint creates multilingual multimedia
• Web sites, animations

17
Web sites

• URLs transmitted in UTF-8 (before the ?)
• FrontPage
• Create and edit web pages in Unicode
• Word
• WYSWYG Web pages
• Save in full or filtered HTML
• IE5
• Display Unicode 3.0 pages

18
Mail and PIM

• Outlook
• New local storage is Unicode
• Contacts, Calendar, Tasks etc.
• Display and message handling all Unicode
• Send/receive mail in any encoding

19
Unicode HTML

• HTML is a companion file format
• Roundtrip all formatting
• Optional HTML Filter cuts file size for
  publishing
• Save to web servers directly
• Roundtrip Unicode data in any encoding
• UTF-8 and UTF-16 are supported too
• HTML is tagged with encoding

20
Unicode e-mail

• Office2000/10 provides fully multilingual email
• HTML mail uses internet standards
• All Unicode content preserved
• Plugs into Outlook, Outlook Express, Exchange
• Use Word to compose replies and new messages
• Send in plain text, RTF, or HTML
• All applications can mail documents as HTML

21
Future Directions

• Help Windows build a worldwide platform
• Ensure system support is useful to app writers
• Unicode 3.0 languages too
• Extend Unicode support to more apps
• Visual Basic Editor and Forms

22
Microsoft Word10 onWhistler

• Demo

23
Questions

• Answers