In the software world, speed and integration of automated processes is expected and demanded. But in the localization world, we still demand human interaction with translation.
LocalyzerQA is one of the ambitious development undertakings we took on at Lingoport in the last few years. It directly takes on and transforms the status quo of cumbersome tasks endured when performing localization QA for software. It’s been a brilliant effort if you don’t mind me bragging for the team, but it hasn’t been an entirely straight line.
Localization QA for software is a particularly convoluted process. After human translation, a linguist will check the application. If a translation change is needed, they will typically file a bug, perhaps including a screenshot and comment of what needs to be changed. Those bugs get reviewed and prioritized. Someone has to track down the string (words) to be retranslated. The application needs to be retested and approved. Often there are many additional interim steps. It’s no wonder localization QA corrections can take months or even years to fix.
Localization QA should be straightforward, and easy to update if it’s going to keep up with the pace of development. A tester needs to be able to:
- Quickly interact within an application and suggest linguistic changes
- Review, approve and implement changes with a click
- After approval, the software repository is automatically updated. No hunting for files and strings
— Boom! It’s done! —
Corrections can go back to a TMS as required, so the same mistake isn’t repeated in the future.
And if Localization QA is made easy, it can fundamentally change the methods and cadence of software localization.
- Using Localyzer, new features in development can be immediately translated in seconds via connections to several machine translation engines, including those in use by TMSs.
- Developers get immediate feedback on how their features work and appear in multiple languages, rather than waiting for a human translation process late in a sprint.
- Translation is always-on, with no concern about minimum human translation charges.
- Translators can edit the Machine Translations right in the context of the application, delivering a higher quality translation included in the QA process.
Remember that the current localization process waits until late in a sprint or series of sprints for source strings to finish any changes – so late in a current development process. Words go out for translation, usually with machine translation being applied in a translation workbench, then edited by human translators. Then they get returned to the development team. Localyzer automates all that, but even with this streamlined process approach, there is a globalization lag from daily and weekly developer progress. Yet machine translation has gotten far more accurate and trainable in the last few years. Changing to an always-on translation process, with QA edits is likely in our future.
Updating translations with LocalyzerQA is so simple, sometimes localization managers and testers have trouble grasping that there are no screenshots or complicated steps involved in updating. There’s no catch like bolt-on proxies or changes to a codebase needed. Lingoport is in a unique position to deliver this because our products already keep track of strings in software repositories. It’s a stride forward in delivering better-globalized products faster and more in tune with software development.
To see how Localyzer and LocalyzerQA works, book a conversation and a demo link here.
To see Localyzer and Localyzer QA working together with a TMS in a continuous environment, view this webinar recording: Software Localization Thriving in a DevOps World (scroll down to view and play).