Speed up Your Translation Processes with Across v7

The new version of the Across Language Server and the Across Translator Edition is now available! We have addressed numerous subject areas in order to improve the user-friendliness, to reduce flow times, and to enable new working styles. 

Get your Across Translator Edition v7 

The first time I wrote about BaccS ("Business accounting system") about two-and-a-half years ago (in edition 265 of the Tool Box Journal), I explained that its sole developer, Eugene (Yauhen) Kuchynski from Belarus, had developed a free cloud-based accounting tool for freelance translators that was generating revenue from Google ad sales. As I said back then, the fact that before Eugene started to work as an English-to-Russian translator he had worked as an accounting software programmer for two years certainly did not hurt his ability to design and create the software. At that point, Eugene had already implemented the ability to import data from the otherwise market-leading TO3000 (v. 11 and earlier), making a switch to his tool much easier.

At the time, he was already gearing up for changing directions and making BaccS a Windows-based application. That was accomplished six months or so later (see edition 274 of the Tool Box Journal), at which point he also put a price tag on the tool. BaccS not only gained a very loyal group of customers (not least through the positive review of Emma Goldsmith) but also a lot of features with additional reporting and project management capabilities.

Now it was time for a change again, resulting in the sale of his company to SDL. If you are surprised by that, let me tell you: I was, too. But I really like it.

I talked to Eugene and SDL's Paul Filkin about this earlier this week, and here is what I learned. Talks about a sale had already started late last year and the actual sale took place early this year. In the months since, the team has been working on integrating BaccS into Trados Studio and GroupShare. But, wait, it's actually not so much a "team" as it is Eugene still working mostly on his own. (I asked about his further plans and he said that there is nothing contractually binding him to SDL, but he seems quite happy where he is right now.)

SDL BaccS comes in four different flavors:

• BaccS Lite is a Trados Studio plugin (you'll need to have the very latest update of the latest service pack for Studio 2019 installed) that allows you to generate and send quotes and invoices as well as some reports from within the Studio interface;
• BaccS Desktop is the Windows-based client application (not only as an integrated plugin) for a single user minus a web client;
• BaccS Team is the same thing, only for multiple users with a shared database; and
• BaccS Team Online is all of that plus a web client.

From those descriptions it should be pretty clear who the user groups are: Lite and Desktop are for freelance translators (Lite really only for those who exclusively work in Trados Studio and don't need any of the project management capabilities -- such as scheduling, integrated email, or user management -- or some of the pre-built integrations -- data export, Dropbox/Google Drive/OneDrive, calendar syncing, etc.), whereas the other two are for larger organizations. This was the first surprise for me, that SDL is trying to also sell BaccS to larger organizations. It'll be interesting to see how this works out. While at the present point BaccS is unlikely to compete with tools like Plunet and XTRF, I can see two possibilities: a) BaccS might grow as a tool and become more attractive as a workflow and accounting tool for present users of those larger tools; or b) there are a lot of smallish organizations that are still struggling with some home-grown programming nightmare for their accounting and workflow needs or that don't really have much of anything to offer to ease project management.

The prices for the editions are $89 (Lite), $360 (Desktop), and approximately $495 (Team). And that was the other thing I was surprised about, that SDL wants their users to pay for the Lite edition. As I said above, unless you really do all your translation in Trados Studio, it's not going to be super-helpful since both its reporting and accounting would cover only a part of your professional activities (those performed in Trados Studio). And maybe I'm wrong, but most translators would be hard-pressed to limit themselves to only one translation environment. It would have seemed a much better idea to give this edition away for free to whet the appetite for a more complete version.

Since the Desktop edition can be used independently of Trados Studio, existing users of the pre-SDL BaccS will continue to be supported and will be offered a free upgrade to the current and the next version of BaccS.

According to Paul and Eugene, an important goal in the roadmap will be the increased ability to communicate with other systems, owned by SDL or not.

Here is what I really like about this deal: I've had countless conversations with developers of translation environments where I'm told that they want to focus on their core competency rather than diluting it into areas where they're not really strong. I get that, and it does sound really good. But on the other hand, it really is counter-intuitive to offer translation environments that are not truly complete environments in that they don't offer all the necessary tools that translators need. SDL has solved this by acquiring outside expertise and technology. This is an important step to becoming a true environment.

PDF Translation for Professionals

Translating PDFs is easier and quicker with TransPDF.

• Available within Memsource and memoQ
• Compatible with all CAT tools.
• Fast log-in for Proz members.

Try it FREE for your next PDF project 

This year is the UN-designated International Year of Indigenous Languages with UNESCO "as the lead UN agency for coordination of the year."

There are three main themes for this year: 1) Support of the revitalization and maintenance of indigenous languages, 2) Access through preserving indigenous languages, creating access for young speakers and collecting and sharing data, and 3) Promotion of the knowledge and values of indigenous peoples and cultures.

At the official website, iyil2019.org, you can get an idea of what kind of conferences, symposia, exhibitions, and other events are being organized by UNESCO throughout this year. Some of the events look super-interesting; some leave one wondering (like a couple of Esperanto events in association with IYIL -- wasn't the goal of Esperanto exactly the opposite of language plurality? Unless, of course, Esperanto is now seen as an indigenous language itself...). At any rate, you would be well-advised to see whether one or several of these events are happening at a location near you. Or you can even get "involved" yourself by offering events.

As translators, we're fascinated by languages, so it's not a hard task to convince us (you) that the survival of indigenous languages is important. I also don't have to repeat to you the dire statistics of the rapid loss of languages that we are experiencing. We know that and we care -- even though we sometimes feel helpless in responding to it. How, though, do we communicate that urgency to others, or how do we at least awaken some interest in the rest of the world?

I would like to share a project with you that I have been working on for the last couple of years and which I think has the potential to make exactly that possible. It can change the conversation about why everyone should care about indigenous languages and their plight from because it's the right thing to if you don't you will lose out.

Translation Insights and Perspectives (TIPs) is a tool that collects data under the following presupposition: All languages are equally competent in expressing any concept, and at the same time uniquely qualified to express additional facets of any concept. If that is so, I (and you and everyone) should have an interest in finding out what those facets are (if I have an interest in the concepts in the first place) and in doing whatever I need to do to make sure language diversity remains at its highest possible level.

TIPs is sponsored by United Bible Societies (the umbrella organization for hundreds of national and territorial Bible societies, not unlike FIT's relationship to national translators' associations), and the text it illuminates via translation is the Christian Bible. Since no text has been translated nearly as often as the Bible (portions of the Bible or the entire text have been translated into more than 3000 languages), it has enormous illumination potential, so to speak.

You can find the tool at tips.translation.bible, where you'll find a very spartan interface with myriad "stories" (explanations and examples of how other languages approach words, concepts, or even grammatical constructs through translation). On the site, you can also find instructions for how to use the tool and submit new stories.

The spartan look belies the fair amount of data that has been assembled so far (though the tool is in its very early stages -- stages, by the way, that by design will not lead to a specific end goal, since the data collection is designed to continue ad infinitum). There are currently approximately 1,900 stories, relating to more than 11,000 passages in the Bible from 370 or so languages. As I said, these are early days, and the database is growing every day.

You can access the data by looking at an alphabetical list of terms, languages, or sources, or by the structure that the Bible's system of verses, chapters, and books provides. Many of the often very short stories illuminate different cultural concepts (such as this one revealing through translation that at least one culture refreshingly does not see infertility as a sickness). Others look at many languages at the same time and how they approach really complex concepts (such as this one on "forgiveness"). To make complex stories like the one on forgiveness more easily digestible, they are then also accompanied by a simplified graphic. Others look at different grammatical concepts and how languages have to deal with their own grammar to express something that the source text really does not provide information for (see this or this). The story that first sparked my curiosity is this one on the Chinese gender of God. If you click on that last link, you can also see a Chinese translation of the story (at the bottom of that page) because our committee felt it was an example of an empowering story for Chinese speakers (as the only language we know of being able to express this concept) that therefore should be accessible in Chinese.

You can also see that every language in a story is hyperlinked to a page where links to the Wikipedia page for that language are provided (many languages are otherwise hard to pinpoint geographically and linguistically) as well as to an online version of the Bible in that language, if applicable.

Empowering indigenous (or majority) language speakers (by saying your language is the only one among the 7000 worldwide that can say it just like that) and empowering everyone via indigenous and other languages (by saying your world has changed because you have been able to look at it through the lens of another, otherwise not-accessible language) are only two of this project's goals. Another clearly is to provide people who are interested in the Bible with a tremendous study tool. And yet another is to point out the power of translation. Translation not only illuminates the readers of the target language but, as this tool shows, potentially everyone else as well.

Please use this tool as much as you find it beneficial, contribute to it as you see fit (we are interested in any language), follow its Twitter account, and help us promote it as a great way to celebrate the International Year of Indigenous Languages and translation.

Spring discount for freelance translators!

memoQ users have more fun!

Save 30% on your new memoQ translator pro license by using this code in the webshop IND_TJ_TP_ND. Offer expires at midnight 31 May 2019 in any time zone. Designed by translators for translators, memoQ translator pro will make your translation work fast and easy.

bit.ly/2VODvnF 

The Wisdom of the Crowds . . . in 2010

Back in 2010, InterpretAmerica and Common Sense Advisory joined forces to produce the first (and so far only) comprehensive survey of the North American interpreting marketplace. One question from that survey has stuck with me over the last decade: "How likely do you think technology is to affect the interpreting profession within the next 10 years?"

• 74.4% of interpreters surveyed said that it was very likely,
• 77.3% of suppliers of interpreting said that it was very likely, and
• 80.72% of end users of interpreting said that it was very likely

A couple of noteworthy points can be extrapolated from these numbers. First, end users had a stronger belief that technology would influence interpreting than interpreters did-a sentiment that holds true today as well. Second, all three groups clearly expected to see technology's influence on interpreting grow in the ensuing years after the survey. We're at the 10-year mark and the wisdom of the crowds in 2010 wasn't wrong. Here are three developments that have caught my eye so far this year.

RSI Goes Big at Facebook's F8 Conference

Remote simultaneous interpretation (RSI) is now gaining wide acceptance throughout the interpreting community. The fact that the conversation has moved on from whether it should even be allowed to a discussion of what the standards should be is an indication that interpreters and professional associations now generally recognize it is not a passing fad or flash in the pan.

Some of the best evidence so far of the acceptance of RSI by end users took place on April 30 and May 1 when RSI technology developer KUDO provided interpreting in 15 different languages for Facebook's annual F8 developers conference. On day two of the event, Facebook allowed KUDO and the interpreters to tell the world what they were up to and what followed was a flood of posts on social media from the interpreter teams for the event based in multiple studios on two different continents. Conference attendees and a lot of curious interpreters the world over were able to tune in live and experience it for themselves.

This isn't the first time that an RSI company has provided services for such a big event. In November 2018, the Swiss company Interprefy provided RSI for a similar number of languages for a "global technology company in California." The difference between these two events was that both the service provider (KUDO) and the end client (Facebook) were openly encouraging people to listen to the interpretation live online. With Silicon Valley juggernauts like Facebook jumping on the RSI train, you can bet there will be more multilingual events like this to come. So, if you haven't considered adding RSI to your service offerings, now is the time.

Speech-to-Speech Translation for Web and Live Conferences

What makes these offerings different from what Google Translate and Microsoft Translator already do? That's a great question.

TranslateLive has several modes for different use cases. The one that stands out the most to me currently is the ability to create a "room" online that anyone can connect to on any Internet-enabled device and hold a conversation in real time in two different languages. The transcribed text appears on the screen, so you can see if TranslateLive has accurately transcribed what you have said. If it hasn't, you can restate what you are trying to communicate until the transcription is correct. This is a big advantage over in-ear S2ST gadgets. The most common use case for this technology? Son-in-laws who don't speak the language of their wives' parents, who live in far-off countries and children of immigrants who can't speak grandma and grandpa's language fluently.

Wordly, on the other hand, is focusing on providing real-time S2ST for participants who don't speak the main language of the conference and whose preferred language is not included on the list for professional simultaneous interpretation. For example, Turkish or Slovene at a tech conference in Silicon Valley. Wordly transcribes the original speeches in real time and then broadcasts a running written machine translation of the speech in the chosen target languages to participants' Internet-enabled devices. Participants can also don headphones or earbuds and listen to a synthetic voice read out the translation as well. You can see a video of Wordly in action here. While still in its infancy, Wordly has a growing number of paying customers using their service at international conferences.

Services like these raise all sorts of questions. And skeptics are surely justified in some of their concerns. Can human connection truly happen when mediated through S2ST? Can a participant really get something out of a presentation if he or she has stare at a smartphone screen and read furiously to follow along? These are just a couple of questions that come to mind.

Speech-to-speech translation is still far from perfect. There are frequent mistakes in even the strongest of language pairs. Even so, these offerings are seeking to expand access to language services. I think more cross-language communication is a good thing. I don't see these developments encroaching on the existing interpreting market yet.

Google AI Experiments with a New Speech-to-Speech Translation Model

Lastly, let me draw your attention to a significant step forward in the development of S2ST technology announced this month by Google.

Speech-to-speech translation models now come in two flavors: cascade models and direct speech-to-speech models.

The cascade model, which is employed currently in mainstream S2ST services (and in the two startups listed above), can best be described as a mashup of automatic speech recognition (ASR), which transcribes the speech. This text is then fed into a machine translation (MT) engine, the results of which are then put through text-to-speech (TTS) synthesis to produce the spoken translation. Think of it as a mashup of several discrete processes each taking place after the other in cascading form.

The second, which Google highlighted in a recent blog post, is a direct model, which eliminates the steps of converting the original speech to text and of converting the resultant machine translation back into synthesized speech. Eliminating these two steps streamlines the process significantly. They call it an "end-to-end speech-to-speech translation model." That's right. Google is now employing the data found in waveforms to produce spoken machine translations. No text required. They are calling the model...drum roll please...Translatotron! And get this. The new technology also allows for the translated utterance in the target language to approximate the voice of the original speaker. Creepycool.

This model is still extremely experimental, but it holds the potential for significant advancement in S2ST technology in the coming years. In particular, the idea of using wave forms to encode spoken language could open the possibility of encoding intonation, punctuation, cadence and emotion into a machine-translated utterance, something current models can't even begin to do.

So, there you have it, it has been a busy first quarter in the interpreting technology space. Thanks for reading!

Do you have a question about a specific technology? Or would you like to learn more about a specific interpreting platform, interpreter console or supporting technology? Send us an email at inquiry@interpretamerica.com.

Imagine all the world's best dictionaries at your fingertips!

For a fixed monthly fee, on all your devices, integrated in your daily work applications!

WordFinder Unlimited -- One service, 5 applications and more than 500 dictionaries in 26 languages.

WordFinder Unlimited was voted the winner of the Process Innovation Challenge at LocWorld in Warsaw 2018.

I met with István Lengyel the other day to talk about his latest project by the name of BeLazy. Many of you will remember István as one of the owners of Kilgray/memoQ, and some might also have encountered him while he was briefly working for translation provider On Global.

It's that combination plus his academic backgrounds in translation and economics that prompted him and a team of five developers to come up with this latest project.

Though BeLazy has not been officially released -- it will be revealed at LocWorld Portugal in June -- István gave me a run-down of part of the system. It is essentially a middleware (a layer that sits between two systems and allows them to communicate) that sits between a variety of different users and a whole range of translation management systems.

In a way, his system attempts to fix and bring together some of the components of the world of translation in 2019 that are not being served well with technology and processes that were developed 10 (or more) years ago.

For instance, today's microtasks or jobs of 5, 25, or 150 words that are part of a contractual agreement with a client to not fall under a minimum fee, create an enormous overhead for anyone involved with the job, whether it's the translator or any of the project managers. To assign, deliver, and invoice for the job with existing systems often costs a multiple of the actual value of such a job. This is one example where István sees BeLazy reducing the overhead by creating an automated layer between the parties and systems involved. The individual job only has to be checked in by the project manager with a single click to be recorded in the translation management system and then routed to a translator who has set her availability so that where the system interfaces with her, it knows to automatically accept any new job during a certain time frame up to a certain size and number. No need for the translator to have her work continuously interrupted by having to answer emails; everything is recorded for her billing purposes and she can focus on translation. And since BeLazy talks to many translation management systems, the interface is always the same and the availability (the "holy grail for language providers," according to István) can even be shared among different clients if they all use BeLazy as their middleware.

Translators would not have to pay anything, and agencies would have to pay 5% of the value of the job up to a fee of 10 euro (at which point it stays at that price no matter how big the job is).

I think this sounds really interesting. It is system-agnostic and was developed by someone with a good feel for the translation world and with good contacts to many of the companies that develop translation management systems (whether it's by companies like Plunet or XTRF, both of which are already connected, or the dozens and dozens of internally developed TMSs). From my point of view, a possible downside is the reality that things typically don't run in as streamlined a manner as this would require. Jobs get constantly delayed or otherwise shifted, and it's never quite as neat as booking 3000 or however many words a day and then being able to work on exactly that. If I were to do that, I would spend many days twiddling my thumbs (or writing longer Tool Box Journals). Of course, there's also the problem that only a handful of our large variety of clients would be likely to ever use the system. But then, as I said, István is smart and has a good vision, so let's just see what comes of it.

A selection of free webinars, designed to help you enhance your career.

Earn points and receive an SDL Certificate of Achievement!

View resources