A computer journal for translation professionals

Last week I stumbled on this quote from Lewis Carroll's Alice's Adventures in Wonderland:

Take care of the sense, and the sounds will take care of themselves.

Though Carroll was not talking about translation, the more often I read it the more appropriate it seems as a description of what we do as translators. Quite beautiful, I think.

This edition of the Tool Box Journal is long overdue and was supposed to be longer (a review of the new version of memoQ will now be in the next edition), but I've had a rough couple of weeks, so it is what it is.

Even in the roughest of times, however, it's possible to sense love and be thankful. I'm thankful for many things, including your support.

Have the happiest of all New Years. 

Some of the most special days of the year are the "days between the years" (as the good German in me likes to say). The days between Christmas and New Year (or, more correctly, Epiphany on January 6) are special, not only because they are often restful and slow, but also because they allow us to look backward and ahead at the same time.

This might be particularly interesting this year because -- at least in my book -- 2014 was a rich year for the development of translation technology, and 2015 could be just as interesting.

Here -- in no particular order -- are some of those developments: 

Free tools for freelance translators have been around for some time: Wordfast (Classic) was free for a number of years; most of the makers of larger translation environment tools and translation management systems offer free but limited "satellite" editions (back in the day, SDLX was the only exception to that rule by offering a fully functional and free freelance version at some point); and lately most web-based tools offer free access to projects that were created within the corporate editions of their tool. And of course, open-source tools like OmegaT are free by definition.

With Google Translator Toolkit and Wordfast Anywhere, there have also been a couple of self-contained, web-based translation environment tools. Most professional translators do well to avoid GTT (which essentially stopped development a few years ago, anyway), but Wordfast Anywhere is a tool that, while not as comprehensive as others, at least gives the budding translator a good idea of what a translation environment tool is about.

With the official release of Abbyy SmartCAT and MateCat (see Edition 240 and Edition 242 if you have access to the archives) earlier this year, I would venture to say that a new level has been reached. Both are fully capable tools; in fact, both provide unique features that cannot be found in other tools (thus "unique";-). And while I'm a bit surprised that there hasn't been more widespread adoption, especially in the case of SmartCAT, I think that this will slowly change.

The reason why both of the companies behind the tools offer the tools for free? Both are thinking of building up a new ecosystem for translation services. Is that a reason to stay away from these tools? Not as far as I can see.

New additions to SDL Language Cloud Machine Translation

SDL Language Cloud Machine Translation is a cloud based Machine Translation service that is designed to work with SDL Trados Studio 2014.

The latest release now makes it even easier to access machine translation across a range of every day programs. You can now:  

• Use machine translation as you work in Microsoft Outlook, Word, Excel or PowerPoint with the new Microsoft Office Add-ins.
• Access machine translation from within your web browser with the SDL Language Cloud Translator

Both of these great new translator tools allow you to use your own terminology and specific industry trained engines to receive high quality machine translation results.

Learn more about these features for Freelance Translators and Language Service Providers. 

How often have we talked about machine translation in the past in a rather theoretical let's-not-really-touch-it way? Turns out that MT is actually widely used among translators. The folks behind Memsource looked at actual usage data from their user base and figured out that about 50% of all users use MT (see Edition 240 of the Journal). That in itself was not too surprising (in fact, I think it could actually be higher if Memsource projects set up by a project manager didn't allow the disabling of the MT feature). What we need to look at is another number: 98% of those users use Google Translate or Microsoft Bing Translator.

This figure is shocking on a number of levels: Much has been said about confidentiality when using these public resources because you have to submit your source texts when you use them for translation (and possibly even your translations in some configurations). What you as a translator think about this state of affairs is really not so relevant -- what's relevant is what your clients think about it. I have a hard time imagining that half of all translation clients are OK with translators submitting their resources to either Google or Microsoft.

The other point to ponder is that Google Translate and Microsoft Bing Translator are not the engines that the MT community has been pointing us toward for the last few years. They have been talking about customized engines and not the generic kinds that Google and Microsoft offer (note that Microsoft also offers Microsoft Translator Hub, which is the customizable variety of its engine). But that's clearly not what the vast majority is using. Why is that and what is to be done about it?

I think this points toward major trends that we will see in 2015. Companies behind engines such as KantanMT, LetsMT, NICT, PangeaMT, PROMT, Systran, Tauyou,and others have to make themselves more available and visible to translators -- not just to the larger service providers but also individual practitioners. We now know that MT is used and that translators are willing to pay for it -- so why not use a variety that will be more likely to find client buy-in because it's not public and is specifically geared toward the kind of text you're translating?

What do these companies need to do? Continue (or: start) talking to and with translators about their needs. Describe in real terms (rather than hyperbole) what their engines can do in what language and with what kind of input. Design the user interface of their engines in such a manner that they can be used by non-technical users. Price the offering in a transparent and reasonable manner. And, lastly, show more creativity in how translators can use MT suggestions (see below). 

I really hope that 2014 was the watershed year during which we decided to find new and better ways to look at machine translation. As I have written in the Tool Box Journal (Editions 237 and 242 if you have access to the archives) and in virtually every other venue where I regularly publish, we have the opportunity to use machine translation not only within the post-editing paradigm but also in new and productive ways (keywords: interactive auto-suggestions, repair of fuzzy TM matches, automatic repair of poor MT matches, etc.).

And while this may not necessarily appeal to the folks who are mainly focused on machine translation as a fully automated or almost fully automated way to produce translation, it has to appeal to those among us who are eager to explore technology as a means to support the translator as the central actor within the translation paradigm.

I've heard back from a number of tool developers who have started to investigate this possibility further, and I've assured them (and those who haven't yet started their foray in this direction): the users (we) will reward those who help them (us) to be more productive. The technology market at the end of 2014 is as open as it has ever been since Trados really took over in the 1990s, and users are more flexible. We are looking at a paradigm change and need the tools to support it. Please make them available. 

Orsolya Bugár-Buday, Hungary  

It's hard to pin down an actual time for achieving mainstream status when something has been coming for a while, but I think it's not too far-fetched to say that 2014 was the year of arrival for cloud-based translation environment tools. Completely web-based tools such as Memsource, XTM, SmartCAT, MateCAT, Wordfast Anywhere, Smartling, Easyling, Lingotek, and Wordbee are no longer the exception, and most translators (at least those contracting with LSPs) have processed a job or two in one of those tools.

Most of us found out that browser-based interfaces don't have to be wildly inferior and are a lot more process-oriented than they used to be. Yet we also discovered that they are less loaded with the functionality we find in the desktop versions of, say, Trados Studio, memoQ, or Déjà Vu. That's partly because it's not easy to match all the functionality of a desktop tool in a browser-based tool (especially if the whole range of browsers is supposed to be supported equally well), but it also aligns with a goal to make tools that are slim rather than bloated. This results in tools that are significantly much easier to use than their desktop cousins, but I understand that many experienced power-users bemoan the loss of more control through advanced features.

Across, which along with memoQ offers a desktop and a browser-based translation interface, has made an interesting decision for its development: rather than first developing a new feature in the easier-to-develop desktop interface, everything is first developed for the browser to then be ported to the desktop. I foresee other tools following suit in this.

I think we will see desktop interfaces for some time still, but they will eventually fade and we will all be the better for it. This will force developers to make decisions on which features really are important (plus it will eventually be easier to bring complex features to the browser), and the operating-system-independent browsers will finally make all that tired talk about Windows vs. Mac vs. Linux completely irrelevant.

Client attitudes toward the cloud will also change. There are still some clients who require that their data be stored only on computers and servers owned by their vendors (i.e., you) rather than on servers located outside your physical control (i.e., the cloud), but this will only be a temporary hesitation. Of course, it's important to distinguish between publicly accessible data repositories to which you might contribute the client's data and those that can be accessed only by you plus anyone else to whom you give access. Both can be cloud-based, but whereas clients are relatively easily convinced of the latter, the former is a harder sell.

Still, at this point it's important for you to know what your clients' needs are and satisfy them accordingly. 

I don't have actual numbers to back up Smartling's growth among the translation environment tools that translators are using, but I have no doubt that it's been the tool with the strongest growth this past year. Now, Smartling is not something that translators actually go out and purchase -- although it would not surprise me if there were a freelance program at some point comparable to the partner program that is presently in place for LSPs. Instead, it's a tool that is typically implemented by translation buyers and then handed down to the LSP and/or individual translators. Many of Smartling's current clients are relatively young companies that just want to have something in place that "takes care" of translation -- which is how Smartling portrays its tool. Not surprisingly, companies with long-standing experience in translation and localization have not bought into it so far, but there are plenty of the other kind of companies out there so the growth will continue.

I mentioned this year that I was hoping Smartling would become more translator-friendly. It's possible that this has already happened to some degree (I haven't had a look for a while), but I'm hoping that this process will (continue to) happen and that they will hire folks with real hands-on experience in translation.  

Exactly four years ago I wrote this in a list of dreams I had for the advancement of translation technology:

"In the last newsletter I wrote about OpenExchange, SDL Trados' app store. Here developers can create and offer applications that can extend the usefulness of SDL Trados.

"Other vendors: follow suit! Not only will you prevent SDL from building up its market dominance, but you will also help users of your tools. (Plus you might learn quite a bit about what users actually want and need.)"

I'm sorry to say that no other tool vendor has followed suit, and I think it's also fair that now, after four years and relatively slow initial growth, OpenExchange has become a brilliant resource for users of SDL products, in particular Trados Studio users. While the presence of many apps has made the Studio upgrade process a little more frustrating (the app developers are often not as fast with app upgrades to the new Studio versions as their users might want them to be), it creates a fantastic opportunity for Studio users to shape the product they use day-in and day-out into exactly what they want it to be (well, almost). On the other hand, it allows the Trados developers to focus on the core features, knowing that the more esoteric features will be addressed by app developers. When I held webinars last January on features that were missing in tools and then presented the tool vendors with those lists (see Editions 231 and 232 -- there will be another webinar like this next May), Daniel Brockmann from SDL prefaced his responses like this:

"In general, I believe one of the most important points for addressing any gaps in the technology is to offer an extensible and interoperable platform, not "just" a TEnT. So kind of a TEnP. :-)"

He went on to point out apps for most of the features that either already had been developed by third-party developers or would be soon.

For a tool developer, that's a nice position to be in. 

Across Language Server v6: The Most Important New Features at a Glance:

• Across Dashboard: Relevant project information at a glance.
• crossWeb Review Mode: Easy involvement of international subsidiaries in the review process.
• Across Data Cube: Easy collection of indicators.
• Project Management Cockpit: Optimized work process for project managers.

Find out more about Across version 6 at www.across.net/en/support/whats-new/.

Not that this is really so relevant to us as translators, but there is an increasing drive for combining out-of-the-box machine translation with other technologies.

For instance, the Chinese search engine and MT developer Baidu released an app in early 2014 that combined image recognition with Chinese-into-English machine translation. This means that you could hold your smartphone over an object and your phone would tell you the name of the object in both Chinese and English. It's a clever idea and, of course, also the fount of many jokes (see a good collection right here).

In 2013 Microsoft Bing's translation app had already combined optical character recognition and machine translation so you could place your smart phone over a printed menu or other kind of text in a foreign language. The app would internally recognize the characters, decipher the text, and then translate it into the language of your choice.

This year Microsoft released the much-heralded Skype machine interpretation feature ("The Star Trek Universal Translator has arrived") that daisy-chains voice recognition, machine translation, and text-to-speech technologies. (It was actually a little more challenging than just "gluing" these technologies together -- see here for a description.)

On the other side of the coin, we see Google employing not machine translation per se in new development but the techniques that it developed for machine translation to interpret pictures.

As I said, these developments are fascinating, but they're all things that don't need to concern us with considerations like job security. Still, we would be silly to ignore them.

Can you think of a better conversation starter when you want to share about what you actually do as a translator? Next time your conversation partners need translation, they'll know whom to call. 

As a subscriber to the Premium version of this journal you have access to an archive of Premium journals going back to 2007.

You can access the archive right here. This month the user name is toolbox and the password is russianoperas.

New user names and passwords will be announced in future journals.

If you would like to promote this journal by placing a link on your website, I will in turn mention your website in a future edition of the Tool Box Journal. Just paste the code you find here into the HTML code of your webpage, and the little icon that is displayed on that page with a link to my website will be displayed.

If you are subscribed to this journal with more than one email address, it would be great if you could unsubscribe redundant addresses through the links Constant Contact offers below.

Should you be  interested in reprinting one of the articles in this journal for promotional purposes, please contact me for information about pricing.

© 2014 International Writers' Group