A computer journal for translation professionals

January may not be the traditional time to be especially thankful -- but then what time is not right and appropriate to practice thankfulness? Susan Bernofsky is a German-into-English literary translator who recently published a new translation of Kafka's Metamorphosis (Die Verwandlung). Oh, how I loved Kafka many years ago (I still do, though not quite as passionately), but that's not why I'm so happy about the translation. Instead, I'm thrilled at how much media attention it has received here in the US.

Articles in publications like Slate Magazine and The New Yorker have focused almost exclusively on various aspects of the translation process, and many others discussed the translation, the intelligent cover, and the story itself. Who says that the media is only interested in machine translation?

(Of course, I'm grateful for the good share of media interest that also greeted Found in Translation, and I'm also thankful that so many are still reading or re-reading it and letting us know how much they enjoy the book. Thank you so much.)

Speaking of Found in Translation, another book of the same title was published a little while ago, this one with the subtitle In Praise of a Plural World. Actually it's just a long essay by Australian Chinese-into-English translator and author Linda Jaivin, and it's a fantastic read about translation, China, Australia, and the meaning of everything. Well, not quite the latter, but it's very rich as some of the tweeted excerpts in my Twitter stream have demonstrated.

Lastly, I'm exceedingly happy and thankful to have finished a completely new version of my Tool Box for Translators ebook. I still need to input some edits, but you will receive more information in a separate mailing later this week.

In theory, technology can be a great equalizer. In practice, however, access to technology is not equal, and as a result, technology in many cases becomes anything but an equalizer.

For instance, there has been a lot of talk about team translation -- something that makes a lot of sense for a certain kind and a certain size of project. Unfortunately, it's difficult to use technologies like shared translation memories, termbases, or other resources in most translation environment tools unless someone owns a significantly more expensive and significantly more complex-to-use server-enabled edition of the tool. This means that many projects are essentially not directly accessible to groups of single translators; instead, they have to be managed through a larger organization such as a language service provider (or the client itself). While often that's the preferable way of approaching large -- and especially multi-lingual -- projects, there are certainly exceptions to that.

Some tools have been very good about enabling the real-time sharing of resources between peers. Wordfast has been allowing this for a long time, and the web-based translation environment tools (such as XTM, Wordbee, MemSource, and Text United) have a similar model. While not completely peer-to-peer since one translator generally has to purchase a slightly more expensive version, these make it very easy and practical for such workgroups to exist. Even the open-source OmegaT can be used in workgroups.

Now memoQ has joined the fray with an offering on its own. The newly unveiled memoQ cloud offers the following: Any user can purchase access to the cloud for a fee of 130 euros a month. This will give access to a cloud-based server that is already configured and ready to use. (Kilgray stresses that the server will be on safe, non-US territory in Germany and France. Really? I may be a little biased, but I think there is as much potential for NSA(-like) breaches there as in the US.)

Once a project is configured and uploaded to the server, you can invite other users of memoQ Professional (essentially, the regular 620-euros-version of memoQ version) to work with you on the project in realtime. Presently the number of cooperators is still unlimited, but Kilgray has already mentioned that this will be changed to a limited number of users.

It is also possible to purchase licenses for the web-based version of memoQ (WebTrans), which is surprisingly close in functionality to the desktop version (though the powers-that-be at Kilgray are oddly shy about marketing it). Presently it's possible to have up to five users of that version at 80 euros a month per seat. (You can also add access to the online terminology tool qTerm for 200 euros a month.)

The actual project setup in the cloud edition is a little convoluted and could be made easier in my opinion, but it seems that this is something that's already on the radar for the memoQ development team.

You can find other helpful information on the cloud product in two blog posts by Kevin Lossner (here and here) -- the indispensable blog for memoQ users (and others) -- or this webinar recording.

All of this is managed through Kilgray's Language Terminal, which I've discussed in the past (issue 217 in the archives). So far it has been a good and helpful place to convert InDesign files to a previewable XLIFF, back up memoQ projects, store and share memoQ settings, and publish your profiles for networking and job acquisition.

Now the next piece of the puzzle has fallen into place with the memoQ cloud offering -- but this isn't where it stops. István Lengyel, Kilgray's still-acting CEO, gave a short presentation just a couple of weeks ago in the context of the TAUS Translation Technology Showcase Webinars. There he sort of went public with what Kilgray sees as the Language Terminal's future. (Unfortunately, TAUS has not published the video because it had problems with its processing. If you're interested and if it still hasn't been released within a few days, let me know and I can send you a copy.)

To summarize what Language Terminal is going to be in István's own words: nothing short of "the best things that can happen to the language industry." Well, who needs humility if things are going right, huh?

To summarize it in less bombastic terms: Kilgray is envisioning Language Terminal as the industry-wide hub of translation management, a large and more profound marketplace than those that are available now, one that offers deep integration into existing technologies, including both memoQ and other tools. I'll continue to discuss more of the upcoming features when they become available, but it's interesting that Kilgray has finally let the cat out of the bag (so to speak) of what their plans with LT are -- I, and I know many others, had been waiting for this for a while.

Not surprisingly, Kilgray is very eager to bring SDL on board and, also not surprisingly, SDL has so far not permitted Kilgray access to its application programming interface through OpenExchange. From a translation professional's perspective, chances are good that the OpenExchange initiative for SDL Trados and Language Terminal for Kilgray will both have a real impact on the direction of our technology in the future.

Looking for your perfect translation partner?

Purchase a new SDL Trados Studio 2014 Freelance license before February 28th and receive a FREE laptop sleeve, or if you're already a Studio 2014 user you can enter our competition to win a training 'date' with one of SDL's experts!

Learn more... 

A week-and-a-half ago, I did one of the most enjoyable things I've ever done (well, let me rephrase that: one of the most enjoyable work-related things I've ever done!).

As announced in the last Tool Box Journal (and the interim mailing I sent out), I presented a webinar called "Translation Technology -- What's Missing and What Has Gone Wrong?" Many of you sent suggestions on things that are missing in today's translation technology, and many more were collected during the webinar itself. You can view a recording of the webinar right here (access is free since the webinar itself was free as well).

Once the webinar was over, I tried to categorize all the suggestions and send it off to all the technology vendors I am aware of. So far only very few have not responded.

Here are the categories and the specifics that I sent, with a request for comment:

Voice recognition

Many translators use Dragon NaturallySpeaking (DNS) as their preferred way of entering text but encounter problems when working with translation environment tools. For example:

• Automatic capitalization at the beginning of the segment does not work properly.
• Handling of inline codes is poor and stands in the way of dictating in coherent segments.
• Some features such as auto-complete don't work with voice recognition (while it makes sense that it would not work for single words, there is no reason that it should not work for phrases).

We would like to see your willingness to do proper testing with DNS, fix shortcomings, and develop workarounds (and document those) where possible. Also, please develop voice macros for those users who use the Professional version of DNS and make those available.

Access to external resources

There was overwhelming support of implementing a deeper integration of external high-quality and user-definable language resources. Many translators are using IntelliWebSearch but would like to see similar -- and extended -- functionality become an integral part of their immediate translation environment.

This could include tool-tip-like suggestions over phrases and terms when highlighting them or suggestions in a separate pane (without the delay that some tools right now experience when loading that data).

This would also mean that you can't just provide access to generic resources like EuroTermBank, TAUS, or Linguee unless there are advanced ways of pre-filtering the data.

There also needs to be an easy and well-documented way to user-define such queries.

Termbases

Less than half of the participants in the webinar regularly use the termbase feature of their translation environment tool. Here are some of the main reasons:

• Difficult to maintain the terminology in the termbase (that's true for essentially all tools)
• Difficult to enter the terminology into the termbase (this difficulty varies between the different tools)
• Reliance on Java and many problems with that (only refers to MultiTerm)
• Difficult to import existing terminology in various formats, especially if it's terminology beyond a simple glossary-like collection of source and target terms
• Difficult to switch between different terminology concepts of different translation environment tools
• Lack of morphological recognition (this does not refer to OmegaT, Star Transit/TermStar, and Across for a number of languages)

Much to my -- and possibly your -- surprise, terminology was the second-most-responded-to area. Advanced translators know that terminology work is important, but many still feel unable to use it adequately. What can you do to respond to that? One participant asked why it would not be possible to have one terminology resource that all the tools can access, I imagine that your knee-jerk reaction will be: No! But is that maybe a way?

If not, how can we make sure that terminology can be more easily entered, maintained, and used?

Translation memories/corpora

Two issues were brought up in relation to translation memories and corpora:

• Why have other tool vendors not followed Déjà Vu's lead of fixing fuzzy match TM segments with termbase/subsegment/machine translation data? This seems to be something that would be relatively easy to implement and, unlike other features -- including in-context exact matches, autocompletion, and subsegmenting -- where vendors have all followed each other, this helpful feature is still only implemented by only one tool.
• Implement practical and powerful quality assurance, spell-checking, and maintenance tools for translation memories. A large number of folks requested that. (You might want to take a look at Heartsome's TMX Editor as a starting point.)

Other suggestions included easier ways to add parts of segments to the TM on the fly and better ways of compensating for penalties with differences in matches due to inline codes.

Machine translation

Following are requests in two categories.

For translation environment tool developers:

• MT-enabled fixing of fuzzy TM matches (see above)
• Termbase/TM-enabled fixing of machine-translated segments
• Implementation of PROMT as an MT option (does not apply to Déjà Vu and MultiTrans)
• More user-friendly setup of MTs

For MT developers:

• Interactive autocomplete (where the queries to the MT engine change according to the text input of the translation)
• Immediate learning of MT output (this was brought up many times)

General user-friendliness improvements

Maybe not surprisingly, this was the topic with the largest number of responses. Here is one response that mirrors what many others also communicated:

"Can developers be persuaded that, although they are IT experts, we freelance translators have varying degrees of expertise in computing packages and troubleshooting IT problems? Their assumption often seems to be that we understand computer programming as well as translation."

Here are some specific points:

• User profiles that detect what kind of user you are with what kind of usage intent and technical expertise and an according adjustment of the UI
• More flexibility in the design of translation grids in specific and the user interface in general
• Integrated conversions of measurements (applies to some tools more than to others)
• More WYSIWYG formatting capabilities, even if not reflected in source text
• More flexibility when splitting and joining segments, especially in Trados, and not just of bordering segments
• More flexible quality assurance features with more language-specific, pre-defined rules
• While some users like using regex-like commands and rule-building, there needs to be an equivalent to that with a graphical UI, especially when it comes to features that are relevant to everyone (if you don't do that, you'll end up with two classes of users).
• Better QA of your products before releasing them. Users are so tired of buggy releases. Several asked: "Do they ever test the products before they release them?" I know that many of you do -- but virtually no one does enough of it.
• There were calls for more human-led rather than video-based training.

Exchange standards

• TMX needs to be less prone to loss of information during the exchange.
• XLIFF needs to be easier to exchange between different tools and different user groups.
• There needs to be a (XML-based?) standard for the exchange of keyboard shortcuts -- the majority of participants (and translators in general) are using more than one translation environment tool and it's difficult to adjust typing habits.
• There needs to be a better agreement of how words should be counted. There are essentially two strategies for that: a) use a word count that emulates MS Word or b) use GMX-V as an exchange standard.
• QA rules between different tools should be exchangeable.
• There needs to be an exchange protocol for server-based processes (or, as one participant suggested: "We should think about creating an exchange server fit for all").

Other

• We need more ways to work in workgroups without purchasing more expensive licenses.
• Specialized segmentation rules for certain types of docs, e.g., patents
• Several users suggested cloud-based backup features for translation memories and termbases.
• Improvement in user-friendliness and complexity of web-based translation management systems

I cannot wait to share their response with you in the second of the webinar series on Wednesday, the 29th. Of course, it would be great if many of you could register right here and help me to figure out what the next steps need to be.

Reduce your translation volume by streamlining the writing process!

STAR MindReader -- a must for every writer

Re-use of existing content is the underlying principle behind MindReader, a text memory system that locates previously written words and sentences enables the user to easily check content consistency and make any necessary updates. MindReader allows authors to focus on creating new text. MindReader is a powerful tool that enables authors to fully leverage a company's valuable content resources.

For more information, please contact: mindreader@star-group.net 

www.star-group.net

About three years ago I wrote about Crowdin and its CEO with the many (Latin) y's, Sergey Dmytryshyn. Last year I had a chance to meet him personally in his native Ukraine, and I've been wondering ever since what fountain of youth makes adults -- and smart adults, at that -- look like 16-year-olds. (You think that's rude to say? Welcome to my former world, in which I was assumed to be 16 when I had already been married for six years and had two kids!)

Sergey's youthful vigor is reflected not only in his appearance but in his ever-evolving product as well.

Crowdin.net is a (so far) free platform for the translation of a number of software file formats in a crowdsourced environment. [These are now complemented with many documents, graphics, and other formats]. The concept is this: After you register, you can create a project in which you can upload any number of the supported file types and create groups of translators (there is a page where you can invite them). You can also mark your project as open for everyone, in which case others can find your project and send you a message asking you for permission to translate. [Now there is a third option that allows you to order professional translation through -- ahem -- One Hour Translation.] Once the accepted translator logs in, he or she can see the list of files in his or her language combination, open, and start to translate. During the translation, the translator will be shown MT matches from the Google and Microsoft MT and translation memory matches [as you would expect, these come from previous projects; if the option is selected, they can also come from a "Global TM," which you contribute to as well in exchange].

When you translate you can select one of the TM (if available) or MT matches or type in your own translation and then "suggest" it. All this is possible with the help of [now configurable] keyboard shortcuts. When you go back to a segment that you already translated the target field will be empty again but you can see your own translation in the Suggestions field where others now can vote for or against it. A file can be exported at any stage and the translations used will be the ones with the highest votes. As the project owner you will continue to be informed about the project in the form of an RSS feed you can subscribe to.

Interesting? I sure think so.

Premium subscribers can read the much longer article in the archives in issue 163.

Just last month Sergey and his team released a new feature that they had been working on for quite a while. This feature makes it possible to use the Crowdin framework within a what-you-see-is-what-you-get (WYSIWYG) environment for any kind of web application. They call it JIPT or Just In Place Translation.

Now, preparing your project for a JIPT process requires jumping through a few technical hoops (you can find a very high-level overview right here), but you need to keep in mind that this is something that's built for developers and not for the technically challenged freelance translator. Once the application is prepared for translation, the interface that translators encounter is easy to use. You can find an example here on the demo page that Crowdin.net offers, or you can also go to a live page that shows translation being executed right now over here (for the latter you'll have to select "Live" under languages; for either you will have to log in with a Crowdin account or a social media login).

On these pages you'll be able to see translatable strings highlighted with surrounding squares. Upon selection, these will open an edit field in which all the necessary resources (TM, terms, MT, comments, etc.) are displayed so you can translate, edit, or vote on the translation.

I talked to folks at Udemy, which works with professional, paid translators of their own choosing, and to people at PrestaShop, which works with a non-paid crowd, about how their translators like to work in that environment. Since both had just started to use the WYSIWYG model, they really only had a limited amount of feedback, but especially François-Marie of PrestaShop mentioned the frustration that translators ran into when translating in the blind (aka "without context"). This problem has been taken care of with this new model, so it's been readily embraced.

I was interested in how the WYSIWYG translation of dynamic content (stuff that sits in databases and is retrieved only at runtime) works with Crowdin's new JIPT tool -- and neither of my contacts was able to give me a good answer for that. According to Crowdin, it works, but for now we'll just have to take their word for it.

By the way, I have rarely talked to users quite as enthusiastic about the product and the support they receive as those I interviewed about Crowdin (generally good words are expected from users whose contact info is provided by the vendor, but François-Marie and David from Udemy really were "digging" the tool). If you want to actually be part of a team working on Crowdin, you can enroll on PrestaShop's page.

When I first looked at the interface from a translator's perspective, I liked it, but it looked very mouse-heavy. It turned out that the keyboard shortcuts that are offered for the list-based translation also work in the new interface (such as jumping between segments or committing a translation). The problem that I had when using keyboard shortcuts was that it was not always easy to locate the location of a newly opened translatable item on the WYSIWYG interface, therefore defeating the very purpose of the WYSIWYG approach.

Overall, I think that Crowdin is a great example of a company that is driven by highly motivated folks who seemed to have successfully made the transition from a free product primarily geared to the open-source community to a now-paid model for a much larger community. They've achieved that by continuing a development path that wasn't copied from some competitor; instead, they've come up with good, new ideas on their own. 

memoQfest Americas: A 3-day translation technology conference to help you stay ahead of the competition!

27 Feb - 1 Mar, Manhattan Beach, Los Angeles

• 20+ specialized sessions
• case studies, best practices from experts
• memoQ master classes
• introduction to memoQ cloud
• excellent networking opportunities

Click here to join. 

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 changingtimes.

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