A computer newsletter for translation professionals

The story of the Hmong people in the US continues to provide powerful (and heartbreaking) documentation that should be required reading for every interpreter and translator (and anyone who has an interest in intercultural communication).  Anne Fadiman's classic The Spirit Catches You and You Fall Down: A Hmong Child, Her American Doctors, and the Collision of Two Cultures has long been part of the curriculum of many language (and medical) programs, but it received renewed attention a few weeks ago when its protagonist, Lia Lee, died.

At about the same time, another story unfolded between award-winning Hmong writer and interpreter Kao Kalia Yang and the editors of the highly acclaimed radio show Radiolab. It's a raw encounter that may leave you scratching your head and searching for answers to what happened and where exactly communication broke down (that's how I responded, anyway). Listen to the radio show first (about 25 minutes) and then read Kao Kalia's account right here. If nothing else, you'll learn a lot about the fine and fragile lines of cross-cultural communications that connect us day in and day out.

It's hard to imagine that anyone in our industry might have escaped the excitement that continues to build around Found in Translation, the book that Nataly Kelly and I co-authored to raise public awareness of our work as cultural communicators. There are many, many articles and interviews out there, including one in a flagship program of National Public Radio (the great radio refuge here in the US).

There is also a plethora of tweets concerning the book (some of them retweeted by me and some by the book's account, @xl8book). The one that made me the happiest came from @CecileSellier:

Just finished @xl8book and want to thank you both for sharing those wonderful stories - being a #xl8r is awesome!

Yes! Being a translator is awesome, and it's exactly that kind of pride in what we do that we hoped to communicate with our book.

Here was another that also got right to the heart of the matter (by @ChemXlator):

Copy of Found in Translation arrived - well, copies. I'll be giving it for birthdays and Christmas!

That's precisely what we're hoping for. Shower your clients and family with the gift that will finally make them understand how relevant your work is.

If you're anything like me, you've moved beyond apologizing for immature jokes. Here is one of my favorites that I like to share with my daughter's middle school classmates:

Q: Outside the bathroom you're American. What are you inside the bathroom?

A: European.

(Keep reading it aloud -- it'll click eventually!)

Since we're talking about Europe, the European Commission finally and after a long and drawn-out process decided which tool is going to replace Trados as their chosen translation environment tool.

And the answer is...

Can you bear the tension???

It is...

Trados.

Hmm. Well, that's what it is, and while this doesn't provide the market shake-up I had hoped for, it's probably a solid and safe decision on the side of the EC. (How goes the saying: "No one was ever fired for hiring IBM"?) I have tried to talk to SDL about this, but since there are still some papers that need to be signed, they are waiting until this next week to publicly discuss it. A representative of the EC Directorate-General for Translation was kind enough to answer some of my questions.

Q: Why did you decide to go with a commercial tool vendor after the unfortunate experience with Systran and the costly lawsuit?

A: The "unfortunate experience with Systran" (your terms) was not the result of using a commercial product, but of a difference in interpretation about the intellectual property rights of each part after years of collaboration in the development of a product. It was linked to the fact that the Commission was further developing a product for which it had acquired rights. The Commission has been using for many years commercial products without any problem. DGT intends to use the product as provided, thanks to the good APIs provided by SDL.

Q: What was different about this second round of the award process that made you decide for a bidder this time?

A: The reason for not awarding the contract in the previous restricted procedure was that no tender had been found to fulfill the minimum quality criteria. In the negotiated procedure that followed, several tenders improved their products with the result that they fulfilled the compulsory requirements

Q: Was there any consideration that SDL/Trados is the market leader and this will inevitably strengthen their position again? Did their market position speak in their favor?

A: The selection process was exclusively based on the criteria set in the invitations to tender and tender specification. The European Commission is market neutral.

Q: When is the move toward the new Trados Studio system planned?

A: Trados Studio should be gradually deployed in DGT starting mid-January 2013.

Q: Will there be major modifications of the product to fit your needs?

A: In principle, no modification of the product will be requested from the company. On the other hand, integration with DGT's information systems will be necessary, as well as automation of some processes. 

The free Terminotix toolbar enables you to launch queries in any Web resources.

WeBiText: The most powerful and unique free online concordancer on the market.

Massively extract terminology from documents or TMX files with SynchroTerm.

Wonder why we call LogiTerm the Swiss Army knife of CAT tools? Ask for your free trial.

Align your documents with the most powerful alignment tool: AlignFactory. Try the free online version at www.youalign.com. 

I've had the opportunity to give a few talks to fellow translators recently, and every one started with this image: 

Unlike Michael Henry Heim, who is pictured in this image, most of us are not literary translators. Still, he and his life can carry great meaning for all of us. When he died a few weeks ago, there were more obituaries than I could count, and not just in translation-related outlets, unless you want to appropriate the Los Angeles Times, the New York Times, and NPR as such.

The reason why so many took note, even though he was "just a translator," was because he truly was a giant of our profession who translated from many languages and for authors such as Mann, Chekhov, Kundera (from both Czech and French), and Grass. In fact, I had very recently been in communication with Heim regarding Grass as I was researching stories for the Found in Translation book.

Take some time to read through some of the obituaries mentioned above, and you'll learn some amazing things about his life. The one that impressed me more than any other is retold in Susan Bernosky's blog (from which I also borrowed Heim's image). She remembers this:

(...) it is a profound loss, not only for those of us who knew and loved him, but for all the readers of the books he translated from a good dozen languages. He really made his mark all over this planet. It was Mike who got the call when the Czech and Slovak halves of Czechoslovakia were in the middle of their divorce; the Czech government wanted to know what words to use in English to name their new country -- the one we now know as the Czech Republic.

Perhaps more than anyone else, translators create language. And those in power who understand the essence of that power turn to a translator when new language is needed.  

I know some of you have been hoping to read a language-related review of Windows 8 in this newsletter. Well, my new Windows 8 computer is on its way, and you shall read something about four weeks from now (and really, there's no dearth of reviews out there).

Just this morning a reader asked for advice on whether a new computer that he is about to order should have Windows 7 or Windows 8. Unless you have a specific reason to run Windows 8 on a computer (such as clients who want you to test software on Windows 8 or you just REALLY want to use the touch-screen interface), I would strongly advise you to hold off. Windows 7 is reliable, and the programs that we use are typically optimized for it. (In the last newsletter I listed the compatibility of most of the translation environment tools with Windows 8. Two stragglers that reported on their tool's compatibility after the release of the newsletter were Across, which will offer support with the upcoming version 5.5, and Fluency, which provides full support right now.)

And this is interesting for anyone who ever dabbles in translation with anything software-related: The Microsoft Language Portal has been refreshed with 57 Windows-8 languages, so you can make sure you know the new terminology whether you run Windows on your computer or not.

Yes, there is a Trados Studio manual!

Greeted with accolades; joint receiver of the SDL OpenExchange popularity prize.

". . . communicates both to a novice as well as a more advanced user." - Jost Zetzsche

"I am finally taking the most out of Studio 2011" - Alejandro Moreno-Ramos

Find out more at tradosstudiomanual.com.

Kilgray-ian Gábor's answer to why memoQ shows only one machine translation hit when you can select a number of MT providers at one time:

We never expected these beasts to proliferate in such a way.

He went on to explain that in upcoming versions, the handling of MT matches will be more advanced.

He's right, of course. A couple of years ago would anyone have expected that MT would become such a key component in the feature set of most translation environment tools today? I for one wouldn't have.

Why did I ask him? Well, last weekend I did a workshop for the AMTA conference on translation environment tools and MT. This gave me an opportunity to see what is offered by the different tools and discuss with the participants (about two-thirds of whom were translators) whether it makes sense to use MT for certain projects and certain language combinations within a TEnT. And if so, in what way.

First, here are a number of tools with their respective plugins to machine translation engines (note that for the actual use of most of these MT tools you need a license key):

Trados Studio 2011

• Integrated plugins for LanguageWeaver (BeGlobal) SDL MT, Google Translate, and Microsoft Translator.
• Free installable plugins on SDL OpenExchange for itranslate4.eu, Systran, MyMemory, and Google Translate and Microsoft Translator (these two last plugins extend the ability of the out-of-the-box connectors; for instance, this allows you -- in the case of Microsoft Translator -- to translate without a license in exchange for your translation data of the current project).
• You can select various engines at a time and all matches are shown in the order of preference you determine. Exceptions: When selecting itranslate4.eu or MyMemory, only those matches are shown.

memoQ 6

• Integrated plugins for Google Translate, Microsoft Translator, itranslate4.eu, Systran, LetsMT!, Asia Online, and MyMemory.
• You can select various engines at a time but will have only one result returned (which always seems to come from Google Translate).
• MyMemory is treated as a TM resource, even though it's a mixture of TM and MT (see below).

Déjà Vu X2

• Integrated plugins for Google Translate, Microsoft Translator, itranslate4.eu, PROMT, and Systran.
• It's possible to select only one MT engine at a time.
• Concordance searches (searches for only one part of a segment) are supported for all engines.
• Advanced use of machine translation in combination with the assemble feature (see the last newsletter, available for Premium subscribers in the archives).

Wordfast Classic 6

• Integrated plugins for Google Translate, Microsoft Translator, itranslate4.eu, WorldLingo, and MyMemory.
• You can select up to three engines at a time and all matches are shown, including various matches from itranslate4.eu.

Wordfast Pro 3

• Integrated plugins for Google Translate, Microsoft Translator, and WorldLingo.
• You can select several engines at a time and all matches are shown.

Fluency 2011

• Integrated plugins for Google Translate, Microsoft Translator, Systran, and MyMemory.
• You can select one engine at a time and matches are not shown automatically.
• Interface for writing scripts to ease post-editing.

Lingotek

• Integrated plugins for Google Translate, and Microsoft Translator.
• Integratable (don't tell me this is not a word -- it is now!) connectors to SDL LanguageWeaver, SAIC Omnifluent, and Asia Online.
• You can select several engines at a time and all matches are shown.

Multitrans

• Integratable connectors to PROMT, Systran, and MyMemory.

MemSource

• Integrated plugins for Google Translate, Microsoft Translator, Microsoft Translator Hub, and Systran.

OmegaT

• Integrated plugin for Google Translate.

Across

• Integratable connectors to Google Translator, Lucy LT, Reverso, Language Weaver, Moses, and Asia Online.

For clarification's sake:

• MyMemory contains a mixture of large collections of translation memories. If matches from the TM are found, they are displayed. If none are found, machine translation kicks in. It's not possible to say which machine translation engine is being used, since Google Translate, Microsoft Translator, a customized engine, and Systran are all used (in that order of likelihood).
• itranslate4.eu is a European consortium of nine different machine translation vendors: MorphoLogic, Systran, Linguatec, pwn.pl, SkyCode, Amebis, Sunda Systems Oy, PROMT, and Trident MT. Depending on the language combination, you might be able to use one or several of the engines simultaneously.

So, here were some of the questions that we discussed in the workshop (there were many more, but you had to be there for the others...):

When using machine translation suggestions, is it helpful or hurtful to have several matches rather than just one displayed?

It was interesting to see that there were very divergent opinions -- some felt that this would stifle creativity, whereas others liked the idea of having four or five different machine translations displayed (provided, of course, that there was no TM match) from which a choice could be made. Here's what I think: If you are willing to pay for all the matches -- most of them cost a small amount, though when multiplied by four or five it may no longer seem so small -- and if you are able to quickly process a number of suggestions, it seems to me that this could actually help your creativity. Hint for research programs: great topic for a master's or PhD thesis.

Below you can see an example of how Wordfast Classic displays various hits. While the segment itself is not a great example of a text well-suited for machine translation, it's interesting to see that the two statistical machine translation systems that are trained on actual data (Google Translate and Microsoft Translator) get the title of the book right, whereas the rules-based engines (Linguatec, Systran, and Trident MT -- all through itranslate4.eu) "translate" it.

We ran a couple of tests (all very unscientific, of course). We started with a text from the website TripAdvisor translated by a customized engine from LanguageWeaver (BeGlobal) that SDL gave us access to. We then compared it to the same text run through Google Translate. This was an interesting experiment because the TripAdvisor content (user comments) is very uncontrolled, so I had a feeling that Google might actually perform better -- but it turned out that LanguageWeaver was a lot better in the three languages we tested (German, French, and Spanish).

We also examined a text I had translated the week before for a Windows-based software product and compared how the Microsoft Translator engine matched up to Google Translate in the nine languages represented in the workshop. For eight languages, Microsoft performed significantly much better; in only one (Korean) did Google outshine Microsoft. The superior quality of MS's engine was maybe not too surprising, considering that the Microsoft MT engine is highly trained on Microsoft materials.

Still, what these (admittedly very subjective) tests showed is that if you do choose to have an MT program help you in the translation process, you will need to make sure to wisely evaluate which engine to use according to your language combination and on a per-project level.

How can these engines be embedded more deeply and more intelligently into the translation process?

In the last newsletter I wrote about how Déjà Vu has taken the lead on using MT more intelligently in its translation, something that we will see other tools do very soon. At the beginning of this year I wrote a column for the ATA Chronicle in which I listed some other possibilities for how TM and MT technology can be more closely integrated. The scenario that got the most traction during the workshop was to use MT suggestions as a framework that can be corrected on the fly with terminology from the more reliable termbase and translation memory. So, assuming that the MT match gets the syntax right -- which might be the case in many language combinations, unfortunately excluding English>German -- and then "fills in" with the correct terminology, this might in fact be an extremely helpful step.

Who is using MT?

We didn't know the answer to that. Later at the same conference, Don DePalma stated that Common Sense Advisory's data shows that 7.3% of freelancers use MT on every project. I would be more interested in knowing how many use it every once in a while, because the variety of projects most of us work on might obstruct the use of MT for every project (if not all projects). Still, one point that was made very strongly during the workshop is that there are probably many closet MT users -- translators who feel that it's not acceptable to admit to using this technology. This might very well be true. MT has made significant inroads as a productivity tool in many translators' work lives -- but knowing how vast the quality differences between language pairs and text types can be, we'll need to be gentle in assuming that our solutions work for others as well.

Translate apps and work up to 40% faster with Across Language Server v5.3!

You've likely heard that Across v5.3 is the fastest, most stable, and easiest-to-use Across yet.

But have you also heard that Across v5.3 includes the ability to translate mobile apps natively and easily?

This revolutionary ability sets the Across Language Server apart from the field completely!

• Support of App Localization for Android, iOS (iPhone), and BlackBerry
• Pretranslation up to 20% faster 
• Building and exchange of crossTank Packages up to 100% faster

Visit www.across.net/newversion  or download our 60-day free trial version from www.across.net/en/form-testversions-request.aspx.

I've mentioned the Terminotix search toolbar for Word a number of times recently, but it is indeed becoming more and more useful. The latest edition lets you define your own resources and allows you to search in TMX files or MyMemory, TradooIT, and Glosbe. Also, if you have Word with the toolbar running in the background, you can call up access to the resources directly from any other Windows application. Plus, you can directly download it for free from here and don't need to request it by email anymore.

And here's something that cracks me up. Since most of the tools I write about are installed on my computer (at least for some period of time), I get to "enjoy" their footprint more than most. One thing that amuses me greatly is the hustling that goes on among the different technology vendors to gain real estate in MS Word. There are special ribbons for Trados, Déjà Vu X2 (with three rather pitiful commands), Terminotix, all the different MT providers, and on and on. A particularly clever one is Wordfast Classic, which installs itself rather brashly (and slightly annoyingly) above all in the Quick Access toolbar.

Another search tool that many of you will have heard of is WordFinder. For those who have not heard, WordFinder is essentially a search mask that allows you to search through the data of a large number of commercial dictionaries and license access to them.

The good folks behind WordFinder have just released an online version called WordFinder Online. This operates on the same principle, but the dictionary data is now accessible on a server. After installing a small application, a keyboard shortcut gives you access to the dictionary data, which differ according to what package you purchase in your language combination (alternatively, you can also use your mobile devices to gain access). The language combinations are between English and Danish, Finnish, French, German, Italian, Norwegian, Polish, Portuguese, Russian, Spanish, and Swedish, and the corresponding resources differ significantly. Access for some languages is limited to Collins dictionaries' data that in a less extended form (about half the data) is also available for free on Collins; other languages are more blessed. And when you learn that WordFinder is a Swedish tool it won't surprise you that the Scandinavian languages are particularly resource-rich, especially Swedish with access to the data of 14 dictionaries and almost two million terms and phrases.

Matching the lightning-fast speed of development of the Terminotix toolbar mentioned above, the Word macros collection TransTools that I've talked about before is also making huge strides. Version 2.0 has a number of features, most notably the Document Cleaner, a tool that, not unlike CodeZapper, cleans out unwanted tags before importing Word files into a translation environment tool. Good (and free) stuff.

As a subscriber to the Premium version of this newsletter you have access to an archive of Premium newsletters going back to May 2008.

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

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

If you would like to promote this newsletter by placing a link on your website, I will in turn mention your website in a future edition of the Tool Box newsletter. 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.

Here are some websites that mentioned the Tool Box last month:

www.transforum.de

lizettegotranslations.blogspot.com

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

© 2012 International Writers' Group