Examples from outside of Africa: The PAN Localization project in Asia
(Presentation by Amar Gurung, download the presentation here)
Note: This page is based on notes taken during the presentation and may contain errors and unintended misrepresentations of Amar's statements. Please download the original presentation!
Presentation of the Pan Asia Networking (PAN) Localization project http://www.panl10n.net/ (funded by IDRC) and coordinated by the Center for Reesearch in Urdu Language Processing (http://www.crulp.org).
Phase 1 - 7 countries and 9 partners organisations
Main focus was to research the key challenges and solutions to enable ICTs in local language. This included different activities:
(1) Capacity development / Training
- Mentor placement (2-6 months)
- National Training (up to 1 week)
- Regional Training (1-4 weeks)
- Summer School in Asian Language Processing (8 weeks)
A survey showed that respondents found mentor placement and summer school most beneficial, unfortunately those are also the two most difficult to organise.
(2) Language Technology Policy
- The project found it difficult to explain to policy makers the policy changes they would like to so, without being able to demonstrate the tools and software that are involved
- The project considered technology l10n work not only as a means to an end, but also as an enabler policy processes.
- Also organise workshops so that policy makers could share experiences.
Phase II (2007 - 2010)
Increased to 10 countries (18 partners, 14 languages). Three regional research components
End-uses were not involved very much in Phase I. Phase II will test if the tools that were developed are useful and how they will impact end-users.
What has worked:
- Shared passion and understanding of the importance of local language computing
- Strong regional secretariat (technically and administratively)
- Country components have local authority to decide local needs
- Knowledge sharing and transfer between country teams
- Flexibility - Immediate adaptation as required. The project plan had to be changed a number of times to adapt to local needs, and there was flexibility to make these changes.
Network was set up with a strong hub (National University in Pakistan) and spokes who contract only with the hub, not directly with the funder. Why is that good?
- Most of the organisations did not have the admin capacity or lacked technical expertise. There was passion, but capacity was needed. The regional hub could provide this support.
- The proposal also contained some flexibility to deal with local specificities. A needs assessment was conducted initially.
Some things to keep in mind:
- Different levels of expertise among network members (some need more support than others / other types of support)
- Intellectual property rights (IPR) - Sometimes it is more difficult to keep things open than keeping them closed. Many different opinions existed within the network on the IPR of development outputs. One model was to form a consortium of all country components who would hold the copyright. That seemed not feasible since country components could change over time. Finally all country components gave all copyright to IDRC and IDRC in turn gave non-exclusive rights to all outputs to all components.
- Proprietary vs. open source software - PAN project has been open about this. There are lots of differences, between working with proprietary or open source software (pros and cons) but there is no need to make the choice for the users.
- Linguistics (computational) expertise - Some countries do not have linguistic education.
- International standards - It is difficult to actively contribute / participate in international standard setting bodies. There are some countries who do not know who submitted the Unicode for their countries.
- Local content creation
- Language policy and local language computing policy
- Time factor - Things take time. There are no short cuts and the project needs to be institutionalized to make sure that efforts have sustainable benefits.
- Institutionalization of l10n efforts
Comments / Questions
Dawit: From Amar's presentation we see that there is lots to learn from experiences in other continents. It also is apparent that in Africa activities are lagging behind. For example, we have not really considered issues around IPR very much. Another really important factor I think is institutionalization. Often, in Africa, projects are driven by the passion of one person (rather than organisational dedication) and if that person moves on, the efforts end.
Reinhard: IPR is important to consider not only to protect something, but also if you want to make sure your things are open and can be shared. If the IPR is not clear, then many commercial companies will not touch it - to make sure they don't risk infringement.
Dwayne: Being an open source localisation group, we found IPR quite important. In some cases, government starts using terminology that is owned by private companies, which can cut the other way.
< 2/24 #2 Layercake & Ecology | Agenda | 2/24 #4 Localisation in Europe >