Translation Technology Round Table - Interview with Olga Beregovaya

I'm Richard Sykes. I'm the organizer of the Translation Technology Roundtable on behalf of the Localization Institute. And I'm here at LOCWorld #40 in Estoril, Portugal. I've run into Olga Beregovaya. Olga has been a pillar of our roundtable in the past two events that have been in the US. And we are starting one now in Europe, and it's going to be held in Heidelberg on September 19th and 20th. Now, Olga, we have held a roundtable in your own backyard in California and close in Boulder, Colorado, but you are going to come and join us in Heidelberg. And I'd like to hear from you what is compelling about, about coming to Heidelberg when it's so far away.

Apart from the sheer beauty of Heidelberg, there is also— the reason is very simple. From my experience in the industry, and I'm on the supplier side, I work for one of the largest LSPs in the world, the way people talk about the industry and the way people talk about technology, what interests the clients and the buyers here is often quite different from what it would be in the US. So for me, it's very important to have that exposure and have more insights into the perspective, technology perspective and industry perspective from the European, from the European side.

Sure, makes perfect sense. We have some different sessions going on in Heidelberg than we typically do in the US as well. One of them, for example, is standards, and this topic was not voted as being a primary topic by our advisory board in the US. But it has been by our European Advisory Board. What thoughts do you have about standards?

My thoughts about standards are, as much as our industry has been run by standards and we claim that we have standards, and we definitely do have standards, no questions about that, and all the standard committees have been doing an amazing job. However, once you start with the actual implementation, with the actual deployment, you often find that interoperability and portability is not always there. Again, talking about Europe, it happens quite frequently that in Europe, maybe a little bit later than in the US, you start migrating clients from one platform to another for a variety of reasons. And whenever you migrate between the platforms, something in those standards often goes wrong. —like your segmentation beat, metadata beat, or something else. So I do believe that that's a very— especially even more so with introduction of machine translation, when you are passing content onto an engine and then you need to return it back, and this is where standards play a huge role. So I think it's a very important subject, and I don't think we're dedicating enough time. It's kind of a big industry known secret that something is I'm not going to work on the interoperability, but I definitely do think it's a very important subject.

So if standards are of interest to you, whether you agree with Olga or not is not the point. What's interesting is to come to the roundtable and experience the dialogue and to put in your own ideas, interchange ideas and impressions and war stories with other participants. All of our participants are Solution architects or people that roll up their sleeves and get involved in technology or have felt the pain over the years as Olga has. One more subject while we're here talking and then we'll wrap it up. How about translation memory? Do you have thoughts about translation memory?

Passionately so. I do have definitely thoughts about translation memory. The industry yet again had been run on translation memories for a long, long time. And that was pretty much our only way of introducing efficiencies internally within language service providers and modeling our pricing models with our clients. Obviously, with the introduction of MT, that has changed because we added another category, and even more so now with advancements in neural machine translation and introduction of Adaptive MT. Where is it going to take translation memory? It's probably— it's a big discussion subject. Are they going to converge? Yeah. Is translation memory— is NMT going to be so good that it's going to take place of translation memory? We don't know. But it's definitely, given the technicalities, that tech— the level of technical awareness of people in the room at technology roundtables, is definitely something I'm convinced that it's going to be a fruitful discussion. And I know that people in the room will have a lot to say on the subject.

Yeah, I think so too. I know one of our other advisory board members has felt a lot of a push from some of his customers to adopt machine translation, and he's struggling with how to position machine translation versus human-based translation. To meet up with the hype that is around machine translation. He probably has a slightly more conservative view as to the value of machine translation than you do. So we're going to have some interesting discussions about that. Also, as a concluding session, we're going to have a special session. It's an optional add-on. It's no extra charge or anything. Some of the people want to talk about the technical stuff and want to go home. There are, there's a whole cohort of people that are really interested in the positioning and marketing and selling of technologies. So we're going to have a special add-on session where we're just going to discuss the positioning and marketing. So that's at the conclusion of the final day. It's 2 days, September 19th and 20th in Heidelberg, Germany. The registration is available online at the Localization Institute website, and we would love to see you in Heidelberg, and I am really looking forward to some of the discussions that we're going to have.

And thank you, Olga, so much for being here. Most definitely. Thank you so much for having me.