A computer newsletter for translation professionals

To be honest, I was never a big fan of "meaningful" quotes attributed to famous people. They always seemed sort of disjointed and disembodied and didn't seem to do justice to their originator.

Recently, however, I acquired a book by David and Hilary Crystal called Words on Words, a handsome, massive volume with "quotations about language and languages." And I have to admit that I have enjoyed reading through them -- in fact, I've republished quite a few on Twitter since. (I never said I wouldn't change.)

One quote by T.S. Eliot struck a particularly appropriate chord:

For last year's words belong to last year's language
And next year's words await another voice.

It immediately reminded me of our industry's very mundane ongoing discussions about aging data and increasingly obsolete translation memories. I felt a bit ashamed of that pettiness when I read the quote in the context of the stunning original poem. But it still resonates compellingly for our changing world of translation. 

These last few couple of weeks have been both exhausting and exhilarating for me -- exhausting because there was "way more than I could handle," and exhilarating because I had the sense of catching a vision for the future of our industry, especially how our (translators') work environment will develop with an exciting mixture of translation memory, terminology, and machine translation. I articulated this new excitement in my ATA Chronicle column and for Multilingual Magazine, so make sure to read those articles when you have a chance. (In case you're not a subscriber to MultiLingual, you can get a free subscription right here with the code d90jst.)

I was so giddy about it that I forwarded the ATA Chronicle article to all the tool developers in hopes of either hearing that they're already looking at some of the developments I was talking about or, if not, to nudge them to forge ahead in that direction.

I'm happy to report that quite a few wrote back confirming that they are indeed exploring more intelligent ways of combining translation memory, machine translation, and terminology technology. That's good news.

While I had their attention, I asked another question I've been pondering. The newly developed HTML5 standard, though not yet completed, is already widely used on the web and is starting to arrive in our inboxes in the form of projects. Though many of the underlying principles are similar or identical to the previous HTML standard as far as translation is concerned, there are some exceptions, as described in this blog entry. Because I'm not an HTML developer myself, I felt it prudent to have some kind of "official" statement as to how safe it is to translate HTML5 files in the different translation environment tools.

Here are the various answers I received from tool developers:

• Star Transit (Michael Scholand): Since HTML5 is still in development, we are watching it. We don't foresee any problems in the necessary adjustments. "Normal" HTML5 files can already be processed with the current HTML 4 filter . 
• MadCap Lingo (David Ferriot): Lingo will support HTML5 when the specification is close to being finalized. That's not happening for at least another year, I think.
• MultiTrans (Shirley Coady): Our XLIFF Editor (Prism version) provides full support for the translation of HTML5.
• memoQ (István Lengyel): For the time being we'll just wait and see. I think [that some of the HTML5 specific tags] is already something you can do in the XML filter.
• Wordfast (Kristyna Marrero): We will support HTML5 in Wordfast Classic in the summer of 2012.
• Text United (Marek Piorkowski): We are working on this and the filter should be implemented on the Text United platform next month.
• SDL Trados (Daniel Brockmann): This will require rather minor changes as far as I can see, and we'll look at updating the elements list in the HTML5 filter soon.
• Fluency (Carl Christensen): Our support for HTML5 looks the same as for HTML 4 -- there shouldn't be much of a difference from the translator's perspective. The new HTML tags and attributes will be visible by hovering over the tags in the source window (same as current functionality).

There you have it. With the exception of some very specific properties of the new form of HTML, most existing HTML filters are already able to handle most of it, and the developers are working on even more accurate processing capabilities.

If your tool was not represented in the list above, chances are the developers were contacted but just didn't respond.  

Celebrate 20 years of SDL with 20% off and more

To celebrate SDL's 20 years of innovation we are giving 20% off freelance products and upgrades. But there is more!

Looking for simply Studio 2011? Here is an exclusive discount: SDL Trados Studio 2011 for only €545 / $645! Offer ends 31st March 2012

http://www.sdl.com/promo/toolbox206

Want to learn how to use SDL Trados Studio? Join our Translation Knowledge Series. 

Reader Frank Jensen wrote to me the other day about quality assurance. Here's what he said:

Maybe you could write some more about the incessant plague of Quality Assurance (...)tools that many companies now use. To me it becomes clear that a lot of the dynamics of languages is lost as machine-like (Trados matching) translation and the manic quest for consistency takes over....  In short quality is lost on the altar of consistency...

I'm glad that Frank points this out as a topic worthy of discussion, particularly after a web conference the other day in which I discussed that very same issue. Before I respond to some of Frank's concerns, we need to understand what these quality assurance tools are and, maybe more importantly, what kind of translation we are talking about.

Let's first talk about what those tools are. Quality assurance tools are traditionally third-party products that allow you to verify certain formal quality aspects of a translated text. Very few of those are language-specific (exceptions include a search for inverted Spanish exclamation and question marks or spaces in front of French punctuation). Instead, they look at things such as consistent use of numbers between source and target, consistent spacing, comparisons of relative lengths of source and target segments, consistent use of tags, consistent use of punctuation, and consistent use of terminology. The most well-known third-party tools are QA Distiller and ErrorSpy as well as the free Xbench. Approximately four or five years ago, the translation environment tool makers realized that they themselves could offer many of these quality assurance features as integrated features within their tools. And as so often happens, once one of the vendors started to integrate those features, the others felt they needed to catch up and came up with similar solutions. So while the third-party tools that I mentioned above still offer some features that you will not be able to find right within your tool, I would suggest that you first familiarize yourself with what your tool actually offers. My experience is that, aside from the terminology component, the quality assurance component is the most underused feature set within translation environment tools.

To come back to Frank's concern about quality assurance features robbing some of the translator's creativity, he might very well be right, depending on what you are translating.

With the kind of functional translation that I typically do, consistency is one of the hallmarks of success. If there's no consistency between the documentation, the user interface, and the website for a product, or within the legal or medical documentation that goes along with a product, my clients have every right in the world to complain about it.

Creativity? When it comes to the bulk of my work, I love to be really creative in finding ways to work more efficiently. (I wouldn't write this newsletter if it weren't for that passion.) And, yes, I also love to do things like translate marketing texts where quality assurance tools play a lesser role. But since that is not my bread and butter, I'm very happy that I can rely on my computer to highlight the inconsistencies my tired eyes have overlooked.

What I can see and would understand as a justifiable frustration is when clients require the use of tools for texts that the tools are not suited for. In those cases, I would recommend talking to the project manager in charge and helping him or her understand why this could be detrimental to the success of the project. But, again, for the vast majority of projects, at least some of the features of quality assurance tools are indeed helpful. 

It seems to be increasingly common these days for translators to share their expertise outside of user groups and other forums, packaging it instead into a product that can be marketed to colleagues. I think it's a great trend: it gives us access to much more comprehensive overviews and introductions into topics that might otherwise be difficult, and it reveals a commendable entrepreneurial spirit among translators.

One case in point is a very comprehensive Trados 2011 manual for freelancers written by Swedish translator Mats Linder. Its 280-some pages cover virtually all aspects relevant to the freelance translator. Yes, a spellcheck would have been helpful and/or a native English proofreader wouldn't have hurt either, but it's one of those books that demonstrates the hard-earned knowledge of the author and his determination to communicate it.

I found it particularly helpful that Mats does not confine his manual to the main Trados application, but he points his readers to a good number of plug-ins from SDL's OpenExchange program as well as other third-party products that can be useful in combination with Trados.

He's also not shy about highlighting some of the shortcomings of Trados, such as the processing of PDF files, but overall he is very positive and encouraging and manages to communicate both to a novice as well as a more advanced user. For instance, I learned some details of joining and splitting files and different fuzziness settings that I was not aware of. One additional feature I would like to see more in-depth coverage of in a future edition is more information about file-format-specific processing.

Overall, if you don't feel that you're getting to the more advanced  concepts and features of Trados Studio 2011 (there's also a version of the manual for the 2009 version), you would be well advised to shell out the $45 needed for this book.

Coincidentally, another venture I'd like to profile today is also run by a former Trados manual writer. Eñaut Urrestarazu has launched a new site called Translation Tutorials that contains instructional videos for all kinds of mainstream tools (such as Trados), along with tools that you might not have heard about, such as Xbench, CatCount, and Change Tracker. Eñaut has organized the sites into various categories, many of which are still empty and awaiting your contributions. The majority of the existing videos are free, no-nonsense, and very informative, with only a couple that ask for payment.

I hope that Eñaut's venture is a success. Asking for user participation from the get-go might be the key to that success. Be sure to drop by and check it out.

A tool of a different kind is translationrating.ru. Put together by the multi-talented Konstantin Dranch, it's a website that gives you comprehensive access to the translation industry in Russia. Konstantin developed a scoring mechanism that a team of independent reviewers applied in evaluating the many different translation agencies in Russia. I understand that there has been some criticism about the method and objectivity, but as Konstantin himself states, the point is to provide an overview of the language industry in Russia that you can't find anywhere else. If you click on the link above you will notice that the website is in Russian. It won't surprise you to hear that there will eventually be an English version, but for now you'll have to stick with one static English page that summarizes most of the findings. When I talked to Konstantin this morning, he alluded to a number of other plans for the site, some of which sound very interesting. Naturally I will keep you updated.

I don't know about you, but there are only so many hours -- sometimes it seems like minutes -- that I can stare at a computer screen and type or dictate text. We all need breaks. I know that some of you have software programs to help you exercise or to remind you to exercise regularly, but a physical sloth like me prefers to exercise another muscle -- my risible muscle.

The last couple of weeks I've been taking my breaks between work sessions by reading Mox's Illustrated Guide to Freelance Translation. The thing that really cracks me up about this ongoing comic strip is how author and illustrator Alejandro Moreno-Ramos captures the freelance translation world through the eyes of his characters with spot-on representations of how they view the world. Even when you don't agree with some of the opinions expressed, you realize that this is still part of who we are: a lovable and yet terribly cantankerous bunch of people.

In the unlikely case that you have never encountered any of the Mox comic strips, you can find a treasure trove on Alejandro's blog. But as Jeromobot and I tend to say, it's much easier to enjoy an actual book while resting on the couch than sitting bolt upright and staring at a screen.  

Recharge your translation business with memoQ!

Track changes, term extraction, cascading, TXML and Regex-based text filters, X-translate, Google v2 API integrated, and many other exciting features.

Download the 45 trial version from www.kilgray.com.

MTV, Machine Translation Views, will be a new irregular feature in this newsletter. I envision it as a place where I can collect some of the many things that are happening in and around the development of machine translation. For those who are new to this newsletter, I by no means recommend the use of machine translation for most projects. But it would be silly for a technical translation newsletter to ignore a part of the translation market that already has had an impact on how some of us work and will have an increasing impact on more and more of us.

Here are some of the things that I stumbled on during the last four weeks:

Microsoft has launched the Translator Hub, a system that allows you to build customized machine translation engines. The emphasis of this tool is on language combinations without already existing engines. Microsoft benefits from this because its developers will be able to use the data that you upload for their own purposes as well. For companies like Google and Microsoft, it's relatively easy to build machine translation engines for languages with lots and lots of materials on the Internet and elsewhere, but when it comes to languages with a relatively small Internet presence it becomes much more difficult.

I have downloaded a large corpus for Inuktitut and English to test the Translator Hub system, and I hope to report on my findings in the next newsletter.

itranslate4.eu is the European consortium of nine different machine translation vendors (MorphoLogic, SYSTRAN, Linguatec, pwn.pl, SkyCode, Amebis, Sunda Systems Oy, PROMT, and Trident MT) that gives access to a large number of language combinations and, in most cases, in a number of varieties from the different vendors. You can use it with an online interface like most other online MT vendors, but it also offers a plug-in to Firefox/Mozilla or to a number of chat clients. memoQ is the first translation environment tool that also offers a direct connection (the fact that the project is coordinated by the Hungarian Academy of Sciences might have helped that a little).

As with Google Translate, if you use it through one of the plug-ins you will have to pay per character, but you can get started with a trial of 10,000 characters.

Again, this tool's differentiating factor is that you will not be shown only one translation but many (depending on what languages the participating TM suppliers support: six in the case of DE>EN, three for EN>RU, or two for EN>ZH). Whether this helps your productivity remains to be seen (it would not help mine), but be that as it may, it's still another interesting new concept.

Many of you are familiar with Kirti Vashee's eMpTy Pages blog. Kirti works for machine translation developer Asia Online and is one of the most vocal and articulate spokesmen of the machine translation community. I think you'll appreciate his latest thoughtful post on the compensation of post-editing machine translation. You can find it right here.

I was just reminded of a new search engine called DuckDuckGo. Apparently it's rapidly gaining in popularity after some of Google's recent privacy controversies. The makers of DuckDuckGo have paid heed to Google's struggles and go to great length on their website to explain their efforts to ensure the privacy of your searches.

Now, DuckDuckGo is not necessarily a replacement for Google or Bing. In fact, if it does not find what you're looking for, it gives you links to both Google and Bing. But beyond that, it has certain features that make it a nice addition to those engines.

DuckDuckGo specializes in quick answers. It calls those answers "zero-click info," data that is presented to you on the top of the page without the need to click on something further. For example, try entering "age of Besse Cooper" to receive this supercentenarian's shocking exact age to the day. Or you could enter "weight of egg" and get the average weight of an egg.

DuckDuckGo derives these kinds of functions from WolframAlpha, the search engine that likes to make you feel stupid. But there are also other things that are specific to DuckDuckGo. Say you can't decide on whether to go out on a date tonight or not. Type in "yes or no" and you'll get the randomly generated answer faster than plucking the petals from a daisy. (You can also try things like "CAT or TEnT". Or my son's prom date quandary last week -- "Gabby or Emily." His answer was Emily.)

Boost your review cycle with Transit NXT's new quality report 

With Transit NXT's new quality report function, you can swiftly check translations for terminology, number formats, length restriction compliance, untranslated segments, spaces and more and create comprehensive error reports. 

Improve your feedback to translators and navigate from error to error to speed up your quality assurance processes!

For more information, send an e-mail to productos@star-spain.com.

The fourth edition of the (English) Microsoft Manual of Style was published just a few weeks ago. You might ask why I'm mentioning that in this newsletter, especially because the many non-English versions of Microsoft style guides are available for free on the web. But it is indeed worth a mention in our context because a whole chapter ("Content for a Worldwide Audience") is dedicated to writing for translation- and localization-related purposes. Fortunately, you can find some of that content on the Microsoft blog, including the interesting guidelines for writing for machine translation, but there's plenty of other stuff in the chapter that you can find only in the actual book.

Still, is this helpful for folks like us who "only" translate? I think it is if you look for relationships with clients that go beyond just receiving and returning jobs. One part of our work as translators has to be that of a consultant, especially when working with small and midsize companies who have not been exposed to translation questions. This kind of consulting is not only profitable but obviously very helpful in our later translation work. Being able to present clients with well-written and well-researched materials like those in the Microsoft style manual has to be an asset that speaks in our favor. 

Reader Mike Shields sent a little essay to me the other day, partly as an apology that he had never subscribed to the Premium Edition but also because he thought it might interest some readers. Here are some amusing excerpts:

I once heard of a Japanese IT firm that issued its staff with foam plastic baseball bats with which to belabor their computers in moments of stress. Now, the editor of a British magazine is offering a monthly 100-word slot on aspects of computers that make the blood boil. But come on! A measly 100 words a month when the average computer can boil enough blood in five minutes to start a black pudding factory? We need more. (...)

Then there is the stupid question. Most software has to be downloaded and installed. But do not fear - you have a "wizard" to help you. Only problem is that some way through the process, this helpful wizard will ask a question like "Do you wish to twirgle the thingletrips?", to which my response is approximately "How the...should I know? You're the ...wizard!", accompanied by a deep bubbling in the aorta. I have software I've never used because I simply cannot understand what I'm being asked or why. Do they ever try these things out on non-nerds?

And what about "Help" tabs? They are about as much use as roller skates on a cattle grid! Just as an example, click on Word help, and enter "Search". This is what you might do if you were new to the system and wanted to know how to search your text for something. But what do you get? "Search for help, templates, and additional online content on Microsoft Office online," "Search for black-and-white clip art," "Search Japanese text," "Letter announcing your job search," and many more such fascinating topics. Help? Help!

A similar effect can be found on almost any website. They all have a "Contact us" tab, but just you try it. It will do anything except let you contact them. Worst is the so-called FAQ list. Who asks these questions so frequently? Certainly not me, nor, I guess, you either. Then, if you do find your way past the barbed wire and gin traps to a section that kindly permits you to write an e-mail, you have to pick from a list of subjects, not one of which is what you want. I once needed to ask how to pay a bill, and I wasted an hour wandering helplessly through minefields like this. Then I tried telephoning, only to waste a lot more time trying to pick my way through another minefield of recorded messages. I eventually sent my enquiry by post. You'd think they didn't want you to contact them -- and you might well be right! (...)

So what can we do? We need more space for moans about the dreadful products and services we get from big-name manufacturers. We need campaigns. We need, in short, a voice for users. A lost cause? All causes are lost if nothing is done. So don't just sit there sizzling like a lot of pressure cookers. Moan! Shout! Complain! You never know, it might just work, to the benefit of everyone's blood pressure. Failing that, there's always a baseball bat. 

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 eigensinn.

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:

zorikioz.wordpress.com

www.translatewrite.com

www.speakingoftranslation.com (thanks for the interview and making the recording available)

www.fidusinterpres.com/category/blog (I feel honored to be mentioned)

www.1-800-translate.com/TranslationBlog (thanks to my favorite blogger for the shout-out and the quote) 

© 2012 International Writers' Group