crossMarket brings together translation service providers and customers. For freelance translators, the crossMarket membership also includes the basic or premium variant of the Across Translator Edition, depending on their membership type.

Start now and save up to 50%: www.crossmarket.net

Just last November I reported this:

And just a couple of months ago, I wrote this:

Before you get completely confused by these two enthusiastic but seemingly unrelated shouts of jubilation, let me tell you how they relate (or should relate). Last November, Kilgray (developer of memoQ) was getting ready to open its redesigned Translation Terminal and was planning to simultaneously release the aforementioned plugin. That was made possible because SDL had changed course and opened its API (application programming interface) to everyone, so Kilgray was able to develop the plug-in they had planned to release for a long time. Embarrassingly for Kilgray, the development was a lot harder than it had thought it would be (quote by Kilgray's Gábor Ugray: "HTTPS is tricky; very, very tricky."), so rather than releasing it last December, they needed an extra two months for the development (as far as I know, this is the first time Kilgray has so badly missed one of its milestones). When it was ready they let me know (I had been voicing my frustration about giving them all that fanfare when they didn't live up to the hype in time) and said that it now only had to go through the SDL signing process for it to be published on OpenExchange and seamlessly work with SDL Trados Studio.

Long story short, it never came to this. So far, SDL has not signed the plug-in, which means that it's not published on OpenExchange and that if you download it from Kilgray's site, it works within Trados but gives you a very ominous-looking warning every time you start Trados Studio: 

Now, it's completely OK to select Yes -- everything will be fine, and there is neither a virus nor harmful code loaded, but to say that this message inspires a lot of confidence would probably not be true.

I had a long and very open exchange of emails with SDL about this. I might not do our discussion complete justice, but this is what I learned from it:

All business entities have eventually one goal in mind: to be commercially successful (duh!). For some businesses that means a large degree of openness in regard to their software assets (and potentially their customers' data), some businesses are open to a lesser degree, and for some it means virtually no degree of openness. Most companies are somewhere in the middle. But virtually all companies want to promote their degrees of openness as being better than their competitors (and/or shame their competitors in the process). And that's where the aforementioned charade comes into play. The openness goes only so far as is deemed beneficial to the commercial success, whether that success is derived through reputation among its users, sharing of resources, or sometimes pretending to be open when they're really not.

In this case, I think all of that is true for both SDL and Kilgray to some degree. It was not wise of SDL to open access to their API and then prevent full access to its benefits by not signing the developed plug-ins (the Kilgray plug-in is the only one I'm aware of that this happened to). And I hope they're going to reverse course on that decision. On the other hand, Kilgray is not nearly as open as it could be. Kilgray does not offer a comparably easy way to develop plug-ins as there is for SDL products, and there has been a very strong push to promote server-based workflows that make exchange standards essentially obsolete.

So, here is what I proposed to SDL (and in turn propose to Kilgray): Negotiate a deal where both tools can read each other's server- or cloud-based translation memories, termbases, and translation projects. This would be something that would truly benefit users of both worlds. And while we still understand that you are both business entities whose ultimate goal is to make money rather than "be open," we might thank you with our purchasing power.

I have written again and again in the last few months that developers are finding ever more reasonable and productive ways to integrate machine translation into existing translation workflows (vs. making translators adhere to machine translation output that needs to be post edited). The latest example of this comes from Patrick Porter. You might be familiar with Patrick. About three and a half years ago he released a Trados plugin on SDL's app store OpenExchange called MT Enhanced using Microsoft Translator. This was a very intriguing tool, and here's what I wrote about it back then:

This tool in its form back then does not exist anymore (see the "trial period clause" mentioned above), but Patrick has enhanced his tool in a number of other ways, including the fact that you can now also use it with Google Translate.

So why would you want to use the "MT Enhanced Plugin for Trados Studio" as it's now called rather than the preconfigured connectors to Google's and Microsoft's MT engines (if in fact you want to use them at all)? Because there are some extra and helpful features:

• You can find/replace any client- or product-specific information on a text before sending it to the MT provider to assure greater confidentiality. You can do the same on a translated text returned to you.
• You can send text for machine translation without tags for better results.
• You send the request to the MT engine only once for each segment (resulting in you paying Google or Microsoft only once).
• And you can connect to Microsoft Translator Hub engine where you can train your own MT engines.

Oh, and the tool is not only free, but Patrick has also open-sourced it so you can take the code and develop it to your own liking (well, the "geeks" among you can, we mortals not so much).

But, believe it or not, this is actually not what I was going to write about. Instead, I was going to share with you another project that Patrick has been working on. He calls it "true interactive translation prediction (ITP) in your CAT tool" and I really, really like it.

Here's what it is: He's using a feature of the open-source Moses MT engine that allows him to access some of its phrase table (the place where the engine comes up with all kinds of subsegment and full segment suggestions from which it eventually chooses one to present to the user). He then presents those to the SDL Trados Studio user via a plug-in as suggestions listed alongside the translation and as auto-suggestions depending on your key strokes. It's really quite cool.

I've been saying again and again that this needs to be one of the crucial ways to approach machine translation: by looking deeper into what the machine translation engine "holds." The one suggestion that machine translations make to the end user is very rarely correct, but that doesn't mean they don't have the wherewithal to come up with "a" correct suggestion (provided they've been fed with enough high-quality data and a professional translator can make the appropriate choice).

So, where can you download it? Well, you can't. Patrick wrote this as a conceptual tool and hasn't released it because he knows that the typical translator won't be comfortable with building her own Moses engine on a Linux server. He has also talked to the Slate Desktop team (which is offering a Windows-based version of Moses), but they had other priorities at the moment (such as the release of their own Trados Studio plug-in).

The good thing is that Patrick has already had some other ideas about going a slightly different route to come to a similar goal, and I really encouraged him to invest some time and energy into that.

And you know what I also asked him to do? If he does go that route, I encouraged him to charge for it. I want to pay for the tools that make me more productive, and I don't want to beg someone to do it. Instead, I want to give him a business reason to do so. So there! 

If I were asked to compile a list of cities known for technological innovation and out-of-the box thinking, Geneva, Switzerland, probably wouldn't be on the list. But, unexpectedly, this world capital of diplomacy, home to many UN agencies and a plethora of international organizations, has sparked a tiny technological revolution that is changing how interpreters are hired.

It is no secret in the world of interpreting that the lion's share of interpreting work is done by freelancers, and that the agencies and organizations retaining their services are all basically dipping into the same pool of freelancers to staff their interpreting assignments. The process of contacting interpreters, checking availability and extending options or firm offers is less than efficient and could stand to be notably improved. Just ask any assigning officer or interpreting project manager. Chasing down interpreters and staffing events with telephone and email is cumbersome, time consuming and downright nerve wracking when trying to find qualified professionals at the last minute.

Enter Interlink Pro, a web-based platform for finding and hiring interpreters quickly and efficiently. This platform's goal is to make the process as painless as possible for both the client and the interpreter. The brainchild of two freelance conference interpreters, Claudinei Nunes Da Silva and Jeremy Engel, InterLink Pro went live in Geneva last year, and in less than one month, they had some 200 conference interpreters registered and receiving assignments. But more importantly, they have fundamentally changed how interpreters are hired on the Geneva market and saved them some money in the process.

How does it work?

Put simply, InterLink Pro is a database of interpreters made available to meeting planners and organizations such as the United Nations, World Trade Organization and others, all of which hire interpreters. Each interpreter has a profile listing his/her working languages, accreditations, certifications, specializations, professional domicile, picture and basic contact information. Each individual interpreter is vetted by InterLink Pro before being approved for inclusion in the database.

So here's the big question. What makes this platform different from the multiple directories already available online like those maintained by the International Association of Conference Interpreters (AIIC), the American Translators Association (ATA) or the countless directories maintained by national and regional professional associations?

InterLink Pro is different because it has included a number of features that make the platform a fast, simple way to receive and then accept or reject work offers for the interpreter, while at the same time making it easy for the interpreter organizer to determine interpreter availability, build teams and extend offers. Here's a list of features that stood out to me when I was given a tour of the platform:

• A calendaring function that allows each interpreter to maintain an up-to-date picture of his/her availability
• A search function that allows the interpreter organizer to search only for interpreters with the desired language combination, professional domicile, accreditations and most importantly that are actually available on the day(s) of a given event.
• Customizable privacy settings that allow individual interpreters to control who can see their profile and availability including firm offers, options and downtime
• A private messaging system that allows interpreters to receive and accept or reject job offers by text message and email

InterLink Pro rolled out first in Geneva. However, the company has plans to expand first throughout Europe and then to other continents.

Interpreters have two registration options "Classic," which is free of charge, and "Premium," which costs US$35.00 per month, billed annually. I suggest you check out their website for details about the two different registration levels.

On the client side of the platform, only international organizations, private agencies, NGOs and certified consultant interpreters can have access to the database and extend job offers. There are also two registration options for employers: "Classic," which is free of charge, and "Premium," which costs US$89.00 per month, billed annually.

My Take on InterLink Pro

InterLink Pro is an idea whose time has come. For organizers, the platform streamlines the process of identifying the right interpreters, determining their availability and extending job offers. For interpreters, the platform provides targeted visibility and a way to receive and accept offers in just a few seconds. Convenience and efficiency are two big pluses for organizers and interpreters alike.

The ability to send and receive job offers via SMS text messages is an amazingly convenient feature. Job offers include three buttons "accept," "decline" and "more information." Interpreters can send short messages with the response as well. Each job offer generates a text message, automatic email, and alert on the interpreter's InterLink Pro profile. The interpreter's calendar is also automatically synced once the offer is accepted.

Another aspect I like about InterLink Pro, is that, unlike other online platforms, it does not require interpreters to "bid" on the same job. Organizers must state their terms in the offer and then it is up to the interpreter to accept or negotiate the terms directly, thus avoiding the all too common "race to the bottom" prevalent on platforms that take a reverse auction approach to negotiations.

Since it is a response to the archaic staffing practices of the Geneva market, InterLink Pro's main focus for now is conference interpreting, however including other types of interpreting is on their roadmap for expansion. This will be critical as they seek to scale up and make the platform a tool used worldwide.

Also, the current calendaring function on InterLink Pro does not synch up with existing calendaring software like Microsoft Outlook, Google Calendar and Apple iCal. So if you use the system you do have to maintain two work calendars or use the InterLink Pro calendar exclusively. However, developers are working on an update that will allow calendar synching in the future.

InterLink Pro is still in its infancy, but its business model is promising and its technology platform easy to use. I'll plan on checking in with them again as they expand to new markets or update their platform with new features.  

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.

LangFM Podcast: Interesting Stores at the Intersection of Language and Tech

Listen to LangFM, an interview podcast full of stories of interpreters, translators and other fascinating people who meet at the intersection of language and technology. Find it on iTunes or at www.langfm.audio. 

Many of you saw the announcement that Xerox's multifunctional printers (MFPs -- devices that print, copy, scan, and fax) can now also be used to translate a scanned page and output it in another language. And many of you may have kind of shrugged it off as a gimmick of sorts. But it turns out that it really isn't a stunt. This is a serious business proposal that presents something of a biggish leap in the world of translation. (If I had to choose one word that's jumped out in my work with adolescents in our community in the last couple of years, it's gotta be "-ish".) Why? Because it might be the most strident example yet of how the Internet of Things (IoT) has become a reality when it comes to translation.

Since the Xerox Easy Translator Service project is actually an ABBYY service that acts as a Xerox "Trademark Licensee," I met with a representative of Xerox (Rui Ferreira, Sr. Director/General Manager Global Office Solutions at Xerox) as well as a number of ABBYY representatives to learn about the kind of outcome they hope to achieve with this.

Rui referred to this project as a "watershed moment after two years of development." It's part of the "ConnectKey" infrastructure that seeks to extend MFPs beyond the hardware that currently supports printing, scanning, faxing, and copying. There is an app gallery (unbelievably only supported by Internet Explorer and Chrome -- and I thought those days were over) that contains a relatively small number of apps, yet all can be downloaded right onto ConnectKey-enabled MFPs (and/or a mobile device). These apps support tools like cloud services, including storing a scanned document in the cloud in places like Dropbox, Box, or OneDrive; faxing a document through a cloud fax solution (yes, Rui assured me, faxing actually does still play an important role); and the Xerox Easy Translator Service. You can order through the Easy Translator Service in three different ways: right through the MFP, through an iOS or Android app, or through a portal where I assume ABBYY hopes most people will eventually end up since this is where you go beyond documents that have been processed with OCR (optical character recognition) and can upload any kind of digital format.

But to come back to the MFP: In the promotional materials it shows a paper document going in (i.e., being scanned in), internally OCR'ed and translated, and then output as a translated paper document. The more typical workflow will likely be that it emerges in digital form, so it's scanned, OCR'ed, sent off for translation, and then returned in digital format. The translation can happen via machine translation or various levels of actual translation by translators and, depending on the quality level, by editors. All this can be set on a per-document level. If you use the MFP workflow, you set one default language combination that has to be changed if another language is required. There is an auto-language-detection feature that is theoretically available, but it's been disabled to increase speed. At this point, if you're waiting for a machine-translated page, it takes 45 seconds to a minute.

ABBYY does not want to disclose which machine translation engine it uses, but it assures the user that it has arrangements with the provider of said engine that any data that is uploaded and translated will immediately be discarded once the translation is done. (Note: For transparency's sake, I think they should disclose who provides that information, particularly because there are only a limited number of options to start with.)

So, why did Xerox choose ABBYY to partner with in this endeavor? There is a very easy answer to that: No other company has the combination of OCR capabilities and translation services that ABBYY has with its diverse array of services and technology. It supports close to 200 languages in its OCR capabilities, although the translation solutions include only about 40 languages (Bulgarian, Catalan, Chinese (S/T), Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Haitian Creole, Hmong Daw, Hungarian, Indonesian, Italian, Japanese, Korean, Latvian, Lithuanian, Malay, Maltese, Norwegian, Polish, Portuguese, Romanian, Russian, Slovak, Slovenian, Spanish, Swedish, Turkish, Ukrainian, Vietnamese, and Welsh) that are dictated by the limitations of the underlying MT engine.

As Slator notes, it's not exactly a cheap solution, but I imagine that its price is not really the main consideration for its primary clients, and at some point anything can be negotiated, I guess.

Now, it's easy to pooh-pooh something like this as a gimmick, as I mentioned above, but I think it's a real coup for ABBYY. Not only has it given them a lot of publicity, but there's no doubt that this provides a really easy entry into organizations that would have been much harder to penetrate. In this way, it's just a matter of using a printer and -- presto! -- a working relationship has been created. It's exactly this kind of creativity that is needed for language service providers big and small to create opportunities. Admittedly, ABBYY has a particular set of technologies, but isn't that the very point of creativity -- to find something that fits you and only you?

And for Xerox? (Which, by the way, initiated Unicode alongside Apple many a moon ago -- thank you, Xerox!) Unless ABBYY embarrasses them (which I think is unlikely), this is an incredibly easy and cheap way to suddenly make a dusty old printer look like a star.

(Did I hear "win-win"?)

• Tackling terminology in a new field: a practical case
• 10 users become Memsource Certified Trainers
• Memsource extends API to all users

ABBYY again. But, no, maybe not, since much of my talk last week with a group from SmartCAT was about the fact that they are actually no longer ABBYY. They legally separated last January, and while there are clearly still some ties (the CEO of SmartCAT, Ivan Smolnikov, is also the CEO of ABBYY Language Services), they are working hard to convey the impression (or reality) that they are two separate entities. They are officially moving from Russia to the US later this year, they have changed their website address (from smartcat.pro to smartcat.ai), and they have even changed the product name (from ABBYY SmartCAT to just "SmartCAT").

Though I had really encouraged them to go this route awhile back, I still asked them why they chose to do this now. They answered that their investors felt really strongly about separating translation technology and translation services -- which no doubt speaks for the foresight of the investors (ABBYY still maintains less than 25% of the company, by the way).

If you don't really know what SmartCAT is, I recommend that you start by reading what I wrote some time ago. There you can also find links to yet more articles.

It's obvious that I've been very positive in the past about SmartCAT -- and I still cheer for them (as I also cheer for many of their competitors).

There are some relatively big changes in SmartCAT as a product from when I last wrote about it. The most striking change is the business model. SmartCAT has been free for freelance translators for a long time (aside from charges for the integrated but optional OCR and MT features), but it's now also free for LSPs. No cost for the installation, no cost for project managers, and no cost for translator licenses. According to Ivan, SmartCAT felt that the paid models have become outdated.

So how do they finance the tool's development and maintenance?

They have identified three different revenue streams:

• An optional SLA subscription that gives better access to support (within two hours and around the clock) and allows for customizations (for instance, if you want to attach your own MT system or want a CMS integration). The cost: between $300 and $2,000 per month, but Ivan was quick to add that most of their registered LSP users don't actually use that.
• Paid MT/OCR services. More than 50% of their user base (across all sectors) uses this at some point.
• A transaction fee for any job transaction through SmartCAT's marketplace (the commission is between 5% and 10% of the total transaction).

It's probably the third option that provides the most hope in SmartCAT's business plan. Access to the marketplace is deeply integrated into the translation workflow (including a "first-come-first-serve" feature), and while it's possible to select your "own" translators and assign work to them, SmartCAT's calculation is that the marketplace alternative might be a welcome opportunity for LSPs working in the not-often-encountered language pairs (particularly, and maybe not surprisingly, because the prices for most of the translators in the database are really quite low).

I asked what kind of evaluation the 55,000 translators in the database had to go through. Aside from a recommendation system, SmartCAT is working on some easy-to-use testing tools that LSPs and others can use to, well, test translators with whom they haven't worked. We shall see how that works.

I personally think we don't need more translator databases that boast about their numbers -- we have plenty of ThoZ -- but we do need multinational databases with truly pre-qualified translators. Hold that thought because you might be hearing from me again about that at some point...

Here are some other interesting facts about SmartCAT: It has a surprising number of 50 developers in R&D (but not a single person in marketing right now!).

Also, it remarkably has more registered translation buyers (1,000) than LSPs (600, of which 100 use it on a daily basis). When I asked whether the translation buyers primarily buy from individuals or LSPs, the response was that it about evens out.

I asked what they thought was their biggest distinguishing feature at this point. (I was curious about this, because while it was the OCR of PDFs when they first started, others have implemented that by now, including Wordfast Anywhere and MateCat.) Their answer: the deep integration of the marketplace into the tool, and the fact that you can have as many translators and/or editors work on one open file at a time without having to split it up ahead of time. 

A couple of follow-ups to last month's Tool Box Journal.

First an erratum of sorts. In the Premium edition's Staying Undercover section, I described how to get rid of personally identifiable information in comments and tracked changes in MS Word (2007 and above). I said that you needed to "select File (or the Office button) and then choose Prepare> Inspect Document (...) and then select the first two options." That's not quite correct. In fact, you need to select only the second option (selecting the first option as well actually deletes all your comments and tracked changes).   

Thanks to Marco Rognoni for this correction.

Another reader -- Lea Rennert -- pointed me to a macro that you can use in Word to change the name of the person who added comments and/or tracked changes. Thanks!

I had also mentioned a new training course for OmegaT. Unbeknownst to me (and the creators of the training course), the Alexandria Library site where the training course was located shut down just the day after my newsletter went out. (No hard feelings -- there were perfectly good reasons for that to happen.)

Still, Marco and his team at Qabiria had to scramble to find a new home, which they ended up building themselves. You can now access it right here, and the good news is that the training is only about half the price!

And lastly, a few months ago, I mentioned that Hacene Dramchini and a team of translators and editors had launched an in-depth technical newsletter on MS Office for translators ("Out of Office"). They unfortunately had to cease operations. The good news is that Hacene will still "provide ebooks, macros and technical support." 

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 GoodFriday.

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.

A Japanese version of the Tool Box Journal can be accessed here and you can subscribe to later editions right here. 

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.

© 2016 International Writers' Group