Takeaways from GALA and AMTA 2018
Click here to close
Click here to close
Subscribe here

Takeaways from GALA and AMTA 2018

Takeaways from GALA and AMTA 2018

Takeaways from GALA and AMTA 2018A fascinating thing about curated public events is that no two people experience the event in the same way. Conferences are particularly like this, especially if they involve multiple parallel tracks. Every attendee must choose their own adventure, and inevitably he/she leaves the event with different learnings and takeaways.

It’s in this context that I’d like to share my experience attending both GALA 2018 and AMTA 2018 a few weeks ago.


My participation at GALA (Globalization and Localization Association) 2018 Boston was initiated by Moravia’s involvement in the TAPICC project, a pre-standard initiative that seeks to eliminate waste and friction between systems involved in translation. After a full-day working session and formal presentation on TAPICC, here are my thoughts:

  • Important progress was made on open actions during the working sessions, especially on fleshing-out the data model for describing tasks. Physically getting people into a room together can create a collaboration dynamic that’s difficult to create otherwise.
  • There seems be broad interest in the initiative from a variety of industry participants, including LSPs, customers of localization services, and platform vendors. And it seems that people want to be involved beyond passive observation, which is a positive sign for TAPICC. I talked to several people who were interested in participating in a pilot/reference implementation of the standard.

A recurring theme, both directly (in the case of David DeLong’s opening keynote, “Building Tomorrow’s Workforce in Today’s Economy”) and in several workgroup discussions was how to attract talent to the language industry.

  • It’s becoming conventional wisdom that employees are more productive and motivated when they feel that the work they’re doing is purposeful. People want to be involved in something that makes a difference in the world, and that same principle is true when it comes to attracting talent—prospects need to believe that the work they’re being recruited for is part of something meaningful. It made me wonder: are we doing a good job as an industry in communicating the fundamental importance of what we do? We’re all working to flip the world’s language diversity from being a barrier to being a human asset. That’s exciting and meaningful! But are we making the case loudly enough?
  • I’ve decided that I’m never using the word “millennial” again when referring to people in their 20s and 30s. After hearing it used enough times to try to generalize the motivations of an entire generation—who are as complex and diverse as any other population—it’s starting to sound like laziness. I strongly believe that to understand people, you can’t just consider one of their characteristics (like age); you need to get to know them. End of rant.
  • There’s a natural cycle of balance in the working world in that the number of people retiring always gets replaced by a similar amount of people who are just starting their careers. It’s a system that arguably benefits from structured, managed attention, such as industry collaborating with academia. I engaged in a few workshop discussions on this subject with folks who brought interesting perspectives to the table, including Laura Linares, PhD candidate in Translation Studies (and one of the winners of the GALA 2018 Rising Star Student Essay contest). Laura’s suggestion was that we look at traditional industry/academia constructs such as internships outside the box of specialized skills, and seek to engage more liberal-arts-oriented thinking in the mix. Companies would benefit from the application of fresh thinking to new problems, and students would benefit from being engaged in real-world, high-level challenges. I think she’s on to something.

And lastly, at GALA, it was interesting and valuable to get insight into the collective thinking: how are customers, technology vendors, translators, LSPs, and others envisioning the globalization challenge? I picked up on a few themes:

  • New modalities of content—in the context of the explosion of publicly-shared video and voice-based interfaces—have shifted the demand landscape completely. I counted eight different sessions that dealt with the question of mitigating language barriers for these new beasts, including George Zhao’s sessions on video localization and Don DePalma’s “Globalization Preparedness” presentation about the changing demand landscape.
  • The idea that localization can’t live in a vacuum, and shouldn’t be separate from the process that ideates and creates content was considered radical not that long ago. But you wouldn’t know it today; it’s now being presented as folk wisdom. For me, this represents significant progress. As an industry, it’s something we need to get straight before we can collectively be more valuable to the whole global content process.
  • It’s always great when traditions get publicly challenged, and we’re forced to question why we do the things we do. The relevance of per-word-based pricing models seems to be under a growing shadow of doubt in the context of changing technology and practices, and it was interesting to hear a challenge to the per-seat software licensing model for translation software, under the argument that it’s bad for collaboration. I’m interested in participating in these debates further, and believe they are a sign of a healthy industry.


The Association for Machine Translation in the Americas (AMTA) event was organized into three tracks: Commercial, Research, and Government. I bounced around between them.

  • There’s a lot of general interest in Machine Translation nowadays, but if this conference was any indication, MT is still largely an academic endeavor. Neural MT is generally agreed to be a significant step forward in MT capability, but there were very few established conclusions, and many more open questions and discussions about what that means. When it comes to how these technologies will be made useful, we’re staring at the frontier. Exciting!
  • Several organizations, including Microsoft, Amazon, and Google are making their neural machine translation toolkits available for free, something that’s reflective of MT’s academic origins and consistent with today’s culture of technological collaboration. This lowers the barrier to entry to build an NMT system, right? Well, yes and no. For me, the elephant in the room is still, “Where does the massive amount of data come from that will train these engines?” This may be an instance where smaller organizations will need to collaborate and share data.
  • There were a couple of presentations in the Government track that I found very interesting. The Dragonfly Project is working on translating American Sign Language into voice—an interesting challenge as the medium of communication is so different from spoken or written text, and is multi-modal. Both facial expressions and hand gestures are at play. As all modern person-to-person communication is multi-modal, I believe that cracking this nut has direct relevance to the broader challenge of translating the tsunami of end-user-produced multimedia.


At the end of it all, I came back inspired with ideas, but more importantly questions about what I see as being some of the most challenging and important topics in our industry. If you were there and have additional takeaways, please share!

Top image: GALA