OneSky is a new crowdsourced translation service that web and app developers can sign up for to have their users translate pages of the site into the languages they know.
Users set up their apps and sites through the OneSky web app, which then scans the content and pulls individual strings into the service separately so as not to overload translators.
Once a user provides a translation and it receives enough votes for accuracy, it can be pushed live using the OneSky API.
In an interface fitting of OneSky’s beta status — not terrible, but feeling somewhat unfinished — app owners are able to peruse strings, their translations and votes, and override user-generated translations.
While many of the pages have been polished off with good design and controls, the translation admin pages in particular are still in sore need of the same treatment, as can be seen below.
In a blog post, OneSky customer Stepcase says that translation is a job that’s never finished thanks to constant optimizations to the company’s apps. That’s what makes OneSky perfect for them, says Stepcase’s Hoi Wan.
“It can take a while with our apps as the UI/UX is constantly optimized, therefore it is never complete. So a new language version will be launched when all high activity screens are translated based on analytics measurement. There are always strings that need removing, strings that need adding, strings that require backwards compatibility support. The role of translation platform becomes important at this point. A highly optimized translation platform ensures that translators are not inundated with vast amounts of strings they have to translate or review,” says Wan.
The app goes up against machine translation services, like Google Translate and Bing Translate, that developers might go to in order to save money, despite their often notoriously bad translations. That said, OneSky doesn’t completely eschew these competitors, with the app providing an option to integrate with these services to jump-start the translation process. It’s a way to add placeholder text while humans get to work providing more accurate translations.