Interview with Dwayne Bailey
Continuous Localisation using FOSS tools. Building a fast responsive localisation process using open source tools
Dwayne Bailey will give a talk about Continuous Localisation using FOSS tools. Building a fast responsive localisation process using open source tools at FOSDEM 2017.
Q: Could you briefly introduce yourself?
Hi, I’m Dwayne Bailey. I’ve been involved in FOSS in some way since I installed a copy of Slackware Linux via ftp, email, uudecode and floppy disk around 1994ish. I must have had serious amounts of time on my hand in those days. I run Translate House which focuses on developing Pootle, an open source localisation platform. That after many years of localising FOSS in South Africa and running language projects in South Africa and Africa. I’m passionate that language should not be a barrier to participation in the digital age.
Q: What will your talk be about, exactly? Why this topic?
Continuous Localisation. This is an emerging topic. If you understand Continuous Integration, you are part of the way to understanding what we’re trying to achieve here. Traditional models of localisation are not responsive to rapid development while Continuous Localisation aims to deliver quickly without compromise in translation quality.
Q: What do you hope to accomplish by giving this talk? What do you expect?
To help reduce the fear around Continuous Localisation and hopefully to get some projects adopting some of our ideas.
Q: You have been working in localisation of FOSS since 2000, building a diverse set of tools such as Pootle, the Translate Toolkit, Virtaal and amaGama. What made you start those tools?
The initial motive was to localise, not to build tools. But as we started localising FOSS we realised that there were many barriers for less skilled people to contribute. FOSS was very much a coders domain. So we started building the tools to allow web based localisation, format conversion, simple desktop localisation and a Translation Memory of FOSS projects. Each one was driven by what we saw as a problem in the FOSS localisation space.
Q: Continuous localisation is for translations what continuous integration is for code. Do both practices have the same advantages and challenges? Or are there important differences?
We’ll be exploring the similarities and differences in the talk. Like CI it doesn’t solve all problems and introduces some of its own. But like CI it’s a solution for an emergent reality. And like CI, it becomes important to realise that the dogged belief that your old process doesn’t have it’s own set of problems needs to be broken.
Q: Suppose I’m running a FOSS project and I want to use continuous localisation. Where do I start and how do I change my localisation process? And what are some common pitfalls I have to look out for?
Start first with your application. A translation should not be able to break your application. Projects that tackle this are well on their way to adapting their localisation processes as their application becomes more robust to unintended issues.
Q: To implement continuous localisation you need the translators to deliver their translations quickly and continuously. How do you encourage them to do this, especially when they are volunteers and they do this in their spare time?
There is the risk of translation exhaustion from a process that does not relent. But the biggest encouragement is a system that is more adept at their translations appearing quickly and timeously to their users. It makes it easier to promote your work, see adoption and be encouraged if your translations appear in days or weeks, rather than months or years.