“Find two Latvians, they’ll form three political parties”, goes a saying about the Latvian tendency to oppose each other. Even if they share common values and goals, they’ll find something to disagree over and waste time in petty quarrelling while diverging from the issues that really matter. As a result, they end up with multiple, often contradicting solutions to the same problem and poor reputation among their neighbors.
Similar mentality seems to reign among the translation technology providers. Competition is fierce, each of them promising to deliver the best solutions for content localization. A myriad of supported file formats, smart user management features, sophisticated quality assessment features. Now, it’s the latter that the average user seems to struggle with the most. And even more so if they have to switch from one tool to another. Even though all CAT tools share the same purpose: get content x translated in languages Y, Z, etc., the way each tool goes about helping the user detect and solve translation quality problems is surprisingly different.
Some of them offer quality review using error typology, others don’t. Those who do, may offer the same error models, but digging deeper, the user might realize that the error types and the logic of error scoring differs from tool to tool. Translation buyers get confused, project managers have to continuously adjust the instructions for translators and the translators have yet to learn another quality evaluation procedure. As a result, we all spend way too much time and energy on the form (quality evaluation methods) and forget about the substance (translation quality).
How do we solve it?
We have to standardize. TAUS, as a neutral industry organization, aims at bringing together the translation community and establishing industry-shared best practices and metrics. Together we know more! TAUS has developed and maintains an open DQF API that enables users to apply the same quality evaluation methods across all CAT tools that they might ever have to work with. It’s an ambitious initiative, but we, the TAUS team, strongly believe that it can be accomplished. However, we cannot do it alone. We need the bright minds of the industry to join us, share their struggles and success stories, and commit to unify the ways we approach quality evaluation.
Needless to say, same goes for politics. While I’d like to be loyal to my fellow Latvians and promote the importance of difference in opinions, I see the value in trying to collaborate and using the collective brain to achieve better outcome for the common good.
On 11 April 2018, TAUS will hold the biannual QE Summit, hosted by Microsoft in Dublin, Ireland. This time the main theme will be the exploration of new approaches towards quality evaluation. Gig economy, neural MT development, changing pricing models for post-editing: these and many other aspects drive the necessity for a common agreement on how we handle and measure translation quality.