Lingoproz Live! A passion for language – The mag for language professionals by Lingoproz.co.za

Found on Open Source Release Feed.com on 20 March 2009

Thanks for joining for this interview. To start of with, please tell us a little more about Guillaume Olivrin.

I grew up in France, in a region with many trees, la Creuse. At 20, I decided to pursue my studies in Scotland and after obtaining my Master degree in Artificial Intelligence, I moved on, exploring the world for a new kind of life. And so it is I came to South Africa with my partner, Fiona, who grew up there. There I found a great life style and a passionating topic: bringing Sign Language onto information systems. This requires the combined skills I had developed in vision and robotics and in computational linguistics. I went on learning South African Sign Language and I have been working with Deaf people ever since to make it all happen.

I work at Meraka Institute at the CSIR as a scientific researcher. I’m also enrolled as a PhD student at the University of Stellenbosch. My homepage is at http://www.meraka.org.za/~golivrin.

Working at Meraka, I know you guys have been building, what is known as the NAP portal. Please tell us some more about this and what you goals are for this project.

The initiative aims at empowering South African people across all disability sectors by providing a one stop “shop” for all information related to disability matters. The portal provides facilities to access, for example, information on Assistive Devices, on Training, on Business Opportunities and on Legal Advice. This initiative brings value to the nation as it attempts to break the current statu quo in the disability domain, for instance on the knowledge and on the availability of assistive technologies in South Africa.

It also provides a framework to create a cohesive and coherent vision of the disability domain across all sectors. It is inclusive and representative of all the role players. Anyone -person, organisation, any NGO -partner or not- can join NAP and use it as its platform of predilection to publish news and keep the community updated. The National Accessibility Portal, NAP for short, is now in its final release stage. It is being taken “out of the lab” so it can be run as a successful national service outside the CSIR.

Technically, the project has been designed with accessibility in mind. From the portal one can expect a fully compliant, tested and tried web access which is particularly adapted for voice synthesisers, which includes South African Sign Language as well as alternative access mechanisms such as telephonic access via a Voice Machine and via SMS on cellphones. It has full multilingual support : it has been localized in 4 languages so far ( Zulu, Sesotho sa Leboa, Afrikaans and English ) and the project is looking for volunteers to participate in localizing in the other official languages of South Africa.

I know that you are currently working on ‘sign language for the web’, SLFW, how is that process going, what have been achieved and what lies ahead?

I am indeed active in finding new ways of providing accessibility for the Deaf by mean of sign language, on the Web and also, more generally, on any computer. As a researcher I have surveyed and reviewed various techniques to provide Sign Language inclusiveness in NAP, for example the localization of navigational elements, menus and error messages and specific sign language content management including captions, videos and graphical notations. Many existing technologies (video, 3D, SVG) already make it possible to reach a good inclusion of sign language on computers and the Web.

But most of these techniques haven’t been systematically applied to sign language, there are few examples. Additionally, there is a lot of guidelines and research questions that need implementation and experimental validation. For example, following the Web Content Accessibility Guidelines and using SMIL for the inclusion of sign language on Webpages is technically possible but there must be empirical evidence ‘out there’ to support this recommendation. This is an example of guideline I would like to see backed up by an initiative like NAP for example.

This is why I am gradually taking part in the W3c activities : to take recommendations forward and make our World Wide Web adapted for sign languages. Globally, sign language needs some sort of electronic support or medium, for example in the format of an electronic ‘book’. There is currently a call to extend the Daisy standard to support sign language. Additionally, there is a lot of guidelines and research questions that need implementation and experimental validation.

Finally, I identified the creation and input of sign language on the Web and on computers in general as the key research question to my thesis. Clearly, it is now possible to use cell phones, digital cameras and high resolutions webcams to capture sign language content but this is not a convenient electronic information format to edit and process. What I see lies ahead is a better ‘format’ and ‘model’ of sign languages content.

As far as open source is concerned which project or projects are you currently involved in and please give us a little background.

First of all my platform of choice is Linux and I have grown quite fond of Ubuntu and all the quality OSS software it brings me. The CSIR had a very forward and courageous initiative to encourage staff to learn and migrate to Linux as their primary operating system. For projects such as sign language, we tested existing open-source software, e.g. for video conferencing (Ekiga, Wengophone, Linphone) to see if we could identify and contribute some code and functionality. We also have home-grown software to contribute such as ATP (the Accessible Technology Platform) on which NAP is built or GNApp, a Grid Communication application and whose source code are candidate for public release.

These projects’ repository is at the moment behind our firewall. The intention is to open these up and use an external repository once we have done some necessary homework. I am involved in a successful scenario of joint collaboration on an open-source project with student projects and the University of Stellenbosch. Another form of open-source contribution, I think, involves as much open-source as open-mindness: that is the participation in bigger-picture projects such as the W3c, Daisy Consortium and to contribute ideas and solution to advance their worthy causes.

Thanks Guillaume for joining me for this interview. In closing, please share some of your thoughts on web accessibility, the open source ecosystem and what the future hold.

We are definitely moving in the right direction as OSS technologies are being adopted all around us by our peers. But the lifeblood of OSS are the actual contributions to the community, and maybe even more so, by the local community. There is a lack of involvement and in this domain. Meraka organizes workshops to set the scene and gather the momentum for all key role players to become actively involved in OSS development.

I see from experience that a project successfully becomes open sourced when two or more partners collaborate over it. Small-one-team projects also benefit greatly from external help and peer-reviewed support – a very scientific approach in fact – be they rarely have the immediate need to setup an open-source environment. This kind of decision therefore must be done before the project starts rather once its too late and the team have lost the momentum. Right now, the decision to open source projects in South Africa is the right one because we need to grow a local interest and skill base into our specific activities.

There is, you are right, a direct link between web accessibility and the open source ecosystem. I am thinking for example of Open Source being essential to unblock situations where we become locked-in with proprietary formats and software. Also, open source is a driving force behind browsers to quickly become more compliant and follow recommendations, guidelines and engage in faster release cycles.