How can we serve a global community? Ushahidi is translated into 18 languages and has been downloaded in 159 countries. We’ve grown fast and our community of translators contributed to many great projects from U-shahid to Qiantang River. Ushahidi’s translators need a new engine with fresh coat of paint. We’d love your input as we navigate this complex issue.
Does not Translate
Imagine you are a deployer in an emergency situation. You download Ushahidi or set up a Crowdmap. On the settings page, you select your language default for the deployment. What if the localization is incomplete? How can you change specific fields and easily contribute to the localization while in the middle of the deployment?
This is exactly the situation that occurred with Cyclone Giovanna Crowdmap. While Ushahidi has been translated into French, there were some incomplete fields. The translator worked with our developer Robbie to fix it during the emergency, and provided some insightful feedback into our tool-set and the mind of a translator.
In Istanbul, I met the creator of Mechulogrenci. Based on his feedback and that of our intern, Michelle Lee, we are on a journey to better meet our localization community needs.
Users before software changes
Ushahidi had an in-house tool called Tafsiri, which we used from 2010 through late 2011. Our mobile developers used spreadsheets to translated the mobile applications as Tafsiri was not useful for their needs. We started to develop Transifex to use for translation but ran into some road blocks with arrays (code). As part of this move, we closed down Tafsiri before Transifex was in place and asked translators to use Github as a temporary workaround. Github is not very translator-friendly.
We own that our migration to Github/Transifex from Tafsiri was a failure. While it was easier for developers to add language to arrays (code), it was a huge barrier to entry for deployers and translators. The move to Transifex did not occur as fast as we intended. We’re sorry for this.
Mind of a Translator (Michelle)
To translate a website means to localize it, it is the way to make the website be accepted by local people and the way to spread the information/knowledge, even more so, the way to collect local information.
For translators, the translate tool must be easy to use. Few minutes tutorial is fine, but if people have to spend two hours to learn how to use it, it will not be the proper tool for the translators. In additional, most of the translators probably don’t have a technical background, it means that if the translate tool needs coding skill, then it will keep the translators away.
The tool should has its own database that can help people to translate most part of the content smoothly, and some specific functions like highlighting the word to get the meaning of if, or can be used both online and offline.
I used to help to translate a business website, we used excel to translate EVERY WORD on the website, it’s annoying and wasted lots of time. And actually, the business site was very simple, no maps or video. Therefore, for a powerful and interactive website like Ushahidi, we need a more powerful translate tool.
Localization Research and Needs
Our friends at Mozilla serve large scale open source localization projects. They’ve provided us with some feedback on how to serve both translators and developers.
If only software localization was as easy as Universal Subtitles! (We use this tool for video translation.)
The next steps:
Being global can be a bumpy road. We aim to meet the deployer and translator needs. The software has to work around this mission, but we think we are on the right path now. One of our major goals this year is to have Ushahidi translated into 25 languages. This means some braining on ways to make it easier and sharing what hasn’t worked.
- Localization and Translation research: Do you have any recommendations or feedback on translation tools?
- Best practices resources for open source organizations?
- Translators: Can you help us define the translator requirements to make your participation easier? Add you comments to this blog post or to our Localization project wiki page.
Heather and Michelle