The first presentation describing the Next Generation Localisation scenario, based on a self-configurable, scenario independent, service-oriented architecture (SOA) framework, with distributed and component-based services that are
extensible and accessible, with
Localisation Knowledge as backbone, and a clear open source IP in place. This idea was later developed into the Service-oriented architecture solution (SOLAS) at the University of Limerick, and its IP transferred exclusively to The Rosetta Foundation which initiated an open source project for SOLAS. SOLAS now powers the Translation Commons or TROMMONS on trombones.org, the language services 'dating' site for nonmarket translation and localization.
Enhancing and Operating Video Collaboration with your Network
Next Generation Localization
1. Scientific Committee Meeting, Dublin, 09 October 2008
Next Generation Localisation
Enabling people to interact with digital content, products and services in their own
language, according to their own culture, and according to their own personal needs
Reinhard Schäler (UL-LOC)
2. Hot off today presses
Personalisation
the
Decision to launch result of growing number
of female drivers in the country, based on
study that researched needs and
preferences.
Women's needs different from men's; most
important thing for them: car will be
comfortable and easy to use.
Last year, Iran's Islamic authorities passed
a bill promoting production of a bicycle
especially designed for women: it will
conceal the rider's legs and upper body.
Iranian women are advised to refrain from
cycling, in order to preserve their modesty,
and are banned from riding motorcycles,
except as passengers.
A woman-only taxi service is operating in
major Iranian cities, and male and female
passengers are segregated on buses and
the Tehran underground system.
Last year, Iran Khodro launched "Islamic" The Irish Times
car model, equipped with a navigation MotorsFeature
Wednesday, October 8, 2008, p.4
system to locate Mecca, will be
manufactured in conjunction with Malaysia
and Turkey.
Robert Tait and Noushin Hoseiny, The Guardian,
Tuesday October 7 2008
http://www.guardian.co.uk/world/2008/oct/07/iran.gender
3.
4.
5. LOC – Next Generation Localisation
Unified Model
Personalised Localisation
Digital
Enterprise Localisation
Content
Management
Next Systems
Generation Framework
Localisation
Integrated
Language
Technologies
6. Next Generation Localisation
LOC1 Multilingual Digital Content Guidelines (Chris Exton, Reinhard Schaler)
WP1 Digital Content Production for Localisation – Beyond Locale
WP2 Localisation Knowledge – Capture, Organisation, Use
LOC2 Translation and Adaptation ( Liam Murray, Reinhard Schaler)
WP1 Technology Evaluation – The process perspective
WP2 Technology Evaluation – The user perspective
LOC3 The Next Generation Localisation Factory ( J.J. Collins, Reinhard Schaler)
WP1 Services Descriptor Development (Web Services)
WP2 Workflow Specification: Bulk (Enterprise) Localisation
WP3 Workflow Specification: Personalised Production & Social Networking Content
WP4 Mining Workflow Patterns
WP5 Collaborative Localisation Platform
8. Next Generation Localisation
Digital Digital Digital
Content Content Content
Processing
Production Localisation
The Next Generation Localisation Factory
9. Today’s Localisation Factory
Based on proprietary technology
Static, constraint environment
OK for pre-defined scenarios
Limited, slow configuration
capabilities
Lack of interoperability
Closed standards
Works well
Within large organisations
For large controlled projects
Example: large multinational digital
publishers
10. Case study
Current throughput: 100,000
The Setting language check-ins per month
2 million files per month
Project constraints
4m wordcount software strings
98% of words leverage
30 languages simultaneous release Average time to process a file: 45
13k localisable files seconds
Localisation group in Dublin; 5,000 people
world-wide distributed development team Fully scalable “add-a-box model”
Simship of all 30 languages
Objectives International version testing before
24/7, 100% automated process – no
exceptions US release
Translation in parallel with development Reduced no. of release engineers
Translation begins at code check-in
Translation “on demand” – no more “big
(20->2) resulting in US$20m saving
project” model per year
Positive ROI within 1 year
11. The Next Generation Localisation Factory
Web-based
Self-configurable (scenario Localisation Platform
independent)
Connecting to the
personalisation agenda (ad
hoc rapid-fire workflows)
Distributed
Component-based
Extensible
Accessible
Localisation Knowledge as
backbone
Localisation
Processing
Clear IP Platform Data
Components Communications
Open source (Layers) Management
12. The Next Generation Localisation Factory
Self-configurable – Adaptable - Extensible
ILT Processing ILT
Tagger MT
ILT ILT
Management Leverager
TM
Translation Router
LOC
Assess type of request and select route according to
time, quality and budget requirements - using available knowledge and resources
LOC
Localisation Knowledge Base
Translation Web Services
Translation Web Services
Access and Communication
Access and Communication
Localisation Data Localisation
Request Response
XLIFF+ XLIFF+ XLIFF+ XLIFF+ XLIFF+
Human Localisation Services
DCM DCM DCM DCM
Personalise Adaptation Processing PersPrefs Profiler
13. The Next Generation Localisation Factory
Self-configurable – Adaptable - Extensible
Components in use
Components not in use
14. The Next Generation Localisation Factory
Self-configurable – Adaptable - Extensible
Components in use
Components not in use
15. The Next Generation Localisation Factory
Self-configurable – Adaptable - Extensible
Components in use
Components not in use
16. The Next Generation Localisation Factory
Self-configurable – Adaptable - Extensible
Management
Communication
Data
Processing
17. Management
Requester Processes
Contact Available
Type (industry, government, ngo) Active
Project Workflow
Constraints (time, cost, quality) Manual
Characteristics (s/t languages, Generated
domain, “value”) Mined patterns
Status
Provider
Contact More on this later
Type (commercial, volunteer)
Resources
Human
Linguistic Localisation
Processing
Processing Platform Data
Components Communications
(Layers) Management
18. Communication
Connection between services
Services descriptor
Webservices description
OASIS TransWeb Standard (initial work)
Areas
Services
Data
Processes
Status
Localisation
Processing
Platform Data
Components Communications
(Layers) Management
19. Data
Data conversion/filter at entry and exit points
Open standards-based data container
Localisation Knowledge
More on this later
XLIFF+
Previous translations, alternative translations
Customer, product, domain information
Status information
Translation directives
Max length
Do-not-translates
Style
Do-not-moves Localisation
Processing
Comments Platform Data
Components Communications
… (Layers) Management
20. Processing
ILT
MT
TM
…
DCM
Personalisation
Integrated by SF
Adaptation
…
LOC
Ad hoc rapid workflow configuration
Localisation Knowledge Base
… (e.g. collaborative Localisation Platform)
Localisation
Processing
Platform Data
Components Communications
(Layers) Management
21. LOC use scenario
Collaborative Localisation Platform
Translators without Borders (int’l collaborators,
partners): “Localisation for free” for NGOs
Ashoka
The Premier Organization Promoting Social
Entrepreneurship Worldwide – system changing
solutions for the world’s most urgent social
problems, working with Muhammad Yunus, 2006
Nobel Peace Prize Winner
International League for Human Rights
Founded in 1942 by Roger Baldwin, ILHR is one
of the oldest international human rights
organizations, exposes serious abuses wherever
they exist.
Doctors Without Borders
Doctors Without Borders/Médecins Sans
Frontières (MSF) is an international medical
humanitarian organization working in nearly 60
countries. 1999 Nobel Peace Prize.
“Access to medicine and healthcare is
determined by economics.”
22. Lack of Localisation can seriously damage your health
If
Localisation is needed to provide access to knowledge
Knowledge helps to address
Health care (netdoctor, doctoronline, …)
Lack of
Justice (lawyers, humanrights, …)
localisation
• Finance (bloomberg, business.com, …)
• Information (weather, exchange rates, …)
can kill
• Research (google scholar, libraries, wikipedia, …)
• Then
– Lack of localisation can seriously damage your
• Health
• Freedom
• Prosperity
23. Phased implementation (collaborative platform)
• Phase I (2009) • User involvement
– LOC Convention (01 Oct 08)
• Academics, industry and NGOs
• Phase II (2010)
– Localization World Madison (13-15 Oct 08)
• Half-day workshop
• Phase III (2011)
– Working Group (starting Oct 08)
• Translators without Borders (TWB)
• Phase IV (2012) – Seminars
• Industry (e.g. Microsoft, Symantec, Sun,
Adobe, Facebook)
• Trial runs (TWB)
• Phase I Implementation by March 2008
– Requester – provider negotiation
– File handling, transfer
– Basic project information
Context Localisation, broadly described as the linguistic and cultural adaptation of digital material to the requirements of foreign markets, started in the mid-1980s. Large US-based developers (then known as software developers, now often referred to as ‘ digital content publishers ’ ) wanted to develop new markets for their products in order to increase the return on investment into their original products. In order to sell into the new markets, they needed to localise their digital products. Initially, localisation was a highly specialised and labour-intensive activity. Different types of professionals were involved. These included software engineers, testers, translators, desktop publishing experts, and project managers. Although these professionals used computerised tools for their work, many of the tasks had to be done ‘ by hand ’ . They were very labour-intensive, highly repetitive, prone to errors and took a long time to complete. Today, the leaders in localisation have adapted what we call the ‘ Localisation Factory ’ , a term first used by Tony Jewtushenko from Oracle at the LRC ’ 03 Conference in 2003. Localisation Factories are highly automated. They rely on tools, process and file exchange standards – all covered under the term ‘ Language Resources ’ in the context of IGNITE. In a Localisation Factory, labour-intensive, repetitive, costly and tedious tasks are automated. IGNITE aims to spread the benefits of such a setting to the industry as a whole by providing general access to Language Resources.