A computer newsletter for translation professionals

Last week was World Poetry Day, and I found this hauntingly beautiful poem by the great American poet W.S. Merwin, himself a multi-language translator. I've reproduced it here from Merwin's collection, The Rain in the Trees:

Losing a Language

A breath leaves the sentences and does not come back
yet the old still remember something that they could say

but they know now that such things are no longer believed
and the young have fewer words

many of the things the words were about
no longer exist

the noun for standing in mist by a haunted tree
the verb for I

the children will not repeat
the phrases their parents speak

somebody has persuaded them
that it is better to say everything differently

so that they can be admired somewhere
farther and farther away

where nothing that is here is known
we have little to say to each other

we are wrong and dark
in the eyes of the new owners

the radio is incomprehensible
the day is glass

when there is a voice at the door it is foreign
everywhere instead of a name there is a lie

nobody has seen it happening
nobody remembers

this is what the words were made
to prophesy

here are the extinct feathers
here is the rain we saw

No one understands the value of languages like we do. That's why we carry a great deal of responsibility -- not just for the languages that we translate out of or into, but for the preservation of all languages. Some say that we have an ethical obligation to protect languages. I agree. But I also think there is a very practical obligation that articulates a much more powerful message to the outside world. In Found in Translation, we describe the Inuktitut term for "Internet" -- ikiaqqivik -- a term that originally referred to "what a shaman does when he travels across time and space to find out about living or deceased relatives." The beauty of this term is that it helps all of us, even the vast majority who don't speak Inuktitut, understand our own language and our own world in a more complex and meaningful way. You don't have to be a "do-gooder" to want to save languages; you just need to be interested in how we all can extend the way we view the world through the "eyes" of different languages.

And speaking of Found in Translation, this tweet made me really happy the other day:

Spreading Found in Translation among our attorney and case manager clients. They should know about our world!

This is exactly what he hoped for! Please keep using our book to explain to the rest of the world what we do and why it's important.

Naomi Sutcliffe de Moraes ended her comprehensive two-part review of LogiTerm in the December 07/January 08 ATA Chronicle (the Chronicle is now available in a beautiful new format) with a ringing endorsement of the product. She said: "If I could have only one translation tool, it would be LogiTerm." A strong statement indeed from someone who has also reviewed plenty of other translation products, though her list of publications and academic work reveals that she has a penchant for working with corpora. But more on that and how it relates to LogiTerm later.

You see, it's not that I was unaware of LogiTerm. I've known about it for many years and have mentioned it regularly in this newsletter, my ebook, and other writings, but I've never written in greater detail about it because -- here it comes -- my pathetic German sense of categories stymied me. LogiTerm never seemed to fit in my own little artificially created Weltordnung of translation tools. Here, then, is my attempt to find a place for it -- or better yet, to grow beyond myself.

Terminotix, the Quebecois company that produces LogiTerm and other tools -- such as the alignment tool AlignFactory, the term extraction tool Synchroterm (both of which I've previously talked about much more extensively), and the concordance products WeBiText and TransSearch -- describes itself as a producer of "CAT tools to help translators achieve higher productivity and quality." That's exactly what it is -- a producer of computer-aided translation tools. Whether it also produces TEnTs or translation environment tools, I'm not so sure.

Oops, here I am with my categories again . . .

LogiTerm, presently in version 5.2 with version 6 to appear before the end of the year, is a tool that operates with three different databases: a terminology database, a bitext database, and a full text database.

Let's start with the last. The full text database is monolingual (Terminotix insists on calling it "unilingual"), and it's essentially a set of indexed documents of all kinds of file formats (Office documents, WordPerfect, HTML, text-based PDF files, etc.) that you can search very, very quickly to find monolingual reference information that is relevant to your current project. You simply create a "module" by identifying or creating a folder somewhere on your computer that holds all the files relevant to a particular topic. When you find a new file or the client sends you new reference data, just save it in that folder, click a button on the status bar within LogiTerm, and the new data is added to the search index and can be accessed on the fly.

The bitext database is somewhat similar, but here the data consists of source and target data that you derive from source and target documents of the file formats mentioned above. These are aligned with a very high degree of accuracy (I've praised the superior alignment of AlignFactory, the same engine used in LogiTerm, many times). The data is kept in either XML- or HTML-based bilingual texts. These preserve the text of the originating documents in its consecutive entirety so you can see all of the context of any match that is found when you search in the bitexts or pretranslate new documents. It's possible to intervene manually in the alignment process with the help of an "Alignment Editor," but typically it's not necessary: the rare alignment errors can be fixed once they are discovered, plus the Alignment Editor isn't that great to start with. (It is also possible to convert translation memories in TMX or Trados 2007 .txt format, but the resulting "bitexts" really are TMs in pseudo-bitext format since TMs typically don't preserve the text in natural order.)

The bitext files can be searched within LogiTerm's interface or used to pretranslate documents. But before talking about pretranslation, a word about the searches. Because they are much more advanced than those of other products, you can use most typical Google search operators (such as plus and minus signs, parentheses, and quotation marks) and wildcards as well as fuzzy match searches, searches that involve both languages (such as all matches for term A in source which do not contain term B in target), proximity searches (how close term A is to term B), and preference searches (such as preferring term A in a certain spelling over term B).

Coming back to the pretranslation: while this is initiated from the interface of LogiTerm, it's MS Word (and WordPerfect) where the actual translation process is performed, so it's primarily Word-compatible documents that can be processed. For some formats, however, including PowerPoint, Excel, HTML, and InDesign (.inx), text is automatically extracted and pasted into the Word interface, where it can then be translated and the translation can then later automatically be reinserted into the original files. It sounds a little complicated, and if it looks like a workaround, that's because it is. If your primary translation format is not Word-compatible files (such as .doc(x), .rtf., and .txt files), I wouldn't recommend relying too much on LogiTerm's pretranslation feature.

Once the file is pretranslated, an analysis is displayed with information on how many matches were found. When the file is opened within Word, the matches are color-coded for how perfect or fuzzy they are. Depending on your previous setting, the translations will already have been entered or you can retrieve and edit them with the aid of the LogiTerm toolbar within Word (which was automatically installed.)

The LogiTerm toolbar also gives you a direct query access (concordance searches) to the full text or bitext databases within LogiTerm as well as to terminology, the third of LogiTerm's components.

I was not overly impressed with the terminology component. The terminology records are stored in Word documents where you can enter them manually with the help of individual term templates, or you can choose to add them semi-manually from existing bitexts or documents that you are currently translating (you can also import glossaries in tab-delimited text format, but there is no support for TBX or other exchange formats). The terminology database(s) can also be used in the pretranslation process and beyond to create so-called "core dictionaries." These allow LogiTerm to take morphological features into consideration during the pretranslation and searches -- if your languages are English and/or French.

It's important to note here that LogiTerm is a Canadian product, a fact that is not only apparent in its market distribution (about 75% of its market is in Canada) but also in its language preferences. While all Unicode-supported languages are supported (I did not test bi-directional languages, but I looked at double-byte languages with Japanese and the processing worked fine), not only is the building of the core dictionaries limited to English and French, but the monolingual term extraction from full or bitexts (useful in building glossaries) is as well, though those languages do indeed provide very clean lists of terms.

One more area where the language preference is apparent as soon as you open the program: aside from the main LogiTerm tab that gives you access to your own LogiTerm resources, there are also (removable) tabs to (the paid) TransSearch with mostly Canadian resources; Termium Plus with English, French and Spanish data; Le Grand Dictionnaire Terminologique with Czech and Yoruba data (just kidding!); and ONTERM, the Ontario terminology database. Naturally I discussed this with Jean-François from Terminotix, and he promised that just as the free Terminotix Word toolbar has developed into a truly multilingual search tool (after some prodding), so will LogiTerm with the upcoming version 6.

All this is not to say that LogiTerm is not usable with other languages, though it certainly proudly carries the "Made in Canada/Fabriqué au Canada" label wherever it possibly can. I asked Jean-François about markets outside of Canada, and he was very happy to report that the Northern European market has been growing rapidly (overall the 12-employee company is growing by about 25% a year).

When I asked about his main competitors -- or which tools he is replacing most rapidly --  it took him about .2 milliseconds to name his Canadian rival MultiTrans. This is probably not surprising considering their main geographical markets, plus the fact that both also supply large international organizations with their technology (for instance, LogiTerm can be found at both NATO and the UN). LogiTerm is also used as an alternative for the indexing tool dtSearch.

Many organizations and translators with specific terminology needs (such as those with very large amounts of translated materials that they need to search very quickly to access translated terminology) use LogiTerm alongside another translation environment tool -- and that's the way I would use it myself. Terminotix markets LogiTerm as the Swiss Army knife of CAT tools, and I see where the idea originates, but if you've ever owned one of those knives you'll know that some of its tools are more helpful than others (my favorite is the toothpick!). It's the same with LogiTerm. The full and bitext and search features are awesome, and some of the other features are helpful -- just not as an everyday kind of thing.

The version of LogiTerm that I've talked about in this article is the desktop-based LogiTerm Pro version (which costs CAN $795), but there is also LogiTerm Web (CAN $2,600 for the server license and one user, CAN $795 for each additional user). The web version, of course, is the version used by LSPs and organizations that want to share their data. You can see one such handsome and public example from the University of Ottawa. 

• CoDyT is part of the Wordbee platform and automates every step in the process.
• Assign jobs to preferred language supplier groups, and free up precious PM time.
• Benefit from self-organizing workflows and seamless document versioning.
• Deliver 200% faster with automated assignments and truly collaborative projects.

Fab 1: The Glossary Links database seems to be standing on shaky ground, but as long as it's there, let's enjoy it! You can find 1,400 different glossaries with a category, language, and keyword search (the keyword search looks only for words in the title of the glossary). The combination of search criteria you'll need to find your results seems a little unpredictable, but you'll find it eventually. I found this through a tweet by Riccardo Schiaffino.

Fab 2: If you have ever translated SAP products, well I should probably say: If you have ever translated SAP products between English and German, you know that everything you have ever known about computer-related terminology needs to be forgotten before you even sit down. It's an utterly different world of terminology -- one that I enjoy, mind you, but it makes me very thankful for terminology resources. Over the years, SAP's approach to supporting the external translation community has been a little capricious, but that has just changed with the amazing SAPterm project, a terminology resource with official translations of SAP terminology into 37 standard languages and 150,000 source entries. Here you can also find links to a terminology community and many other resources. This reminded me of a talk I once had with Aiman Copty and Heinz Lueken, then the responsible representatives for translation at Oracle and SAP respectively. SAP had just announced that it was going to be the first ERP company to translate into Vietnamese, something that was still in the planning phase at Oracle. So I asked Aiman and Heinz -- very naïvely, might I add -- whether this would not be a tremendous opportunity for these two fierce competitors to share terminology so that the customers would not be confused with contradictory terminology, a trademark of Oracle and SAP's history. I actually don't remember what their exact words were, but the gist was: "You've got to be kidding!" No, wait, I think they actually didn't say anything, they just gazed at me with serene and very pitying smiles on their faces.

Fab 3: Linguee just released a whole plethora of new language combinations (DE<>ES, DE<>FR, DE<>PT, ES<>FR, ES<>PT, and FR<>PT). If you work in one of these language combinations, this is very, very good news for you as the thousands of translators who use Linguee in the "old" language combinations (EN<>DE, EN<>DE, EN<>FR, EN<>PT, EN<>PT, ES<>PT, EN<>ES) day-in and day-out will tell you. Linguee is the most comprehensive index/bilingual corpus of web data in combination with morphologically "intelligent" dictionary data. It is often very helpful as an answer to your terminology questions. Even more, it often provides a great starting point for finding sites to research (because every entry you find in Linguee comes with information on where to find the website that information comes from).

Some of you will remember that about a year ago (209^th edition) I waxed eloquent about two paid products offered by Linguee: Linguee Premium and Linguee Professional. While these are only available for the EN<>DE language combination directions (and, due to silly MT vendor Babylon-owned patents, are not available in the US!), they are absolutely fantastic because they evaluate the context of the complete sentence when you look for a term and give you very accurate matches. Really, really cool. So cool that I've been considering leaving this country so I can access that service. (That's a lie -- but it really is helpful.) I asked the folks at Linguee how the product has been received, and here is their very cautious answer: "Quite a few translators have subscribed to Linguee Premium and Linguee Professional, but the programs haven't yet become a large source of revenue. In contrast to our free products, though, the subscription programs haven't been available for long, so it's difficult to make an accurate assessment of the numbers at this point."

To me this sounds like the products are not doing so great -- if that's true, then it's purely a matter of communication. If you are a translator in those language combinations, start using Linguee's products. You'll like them. Promise.

Fab 4: To make the "Four Fabs" heading work, here is the fourth item: I've finally gone through the last five years of Tool Box newsletters that are accessible on the web and added a table of contents of sorts. You can view a preview version right here. You can access the fully linked version if you're a subscriber to the Premium edition of this newsletter (thanks!) or by purchasing the Tool Box ebook (a great reference tool that even I as the author use regularly!). 

Here is a very high-level summary of what XTM is and does in a couple of sentences: XTM is a full-fledged, cloud-based, SaaS (software as a service) translation environment tool with robust workflow components that strongly emphasizes translation exchange formats. While it was originally designed as a tool for larger companies, it shifted more of its attention to the freelance and smallish(er) LSP market about two-and-a-half years ago when it launched XTM Cloud. That was exactly when I last wrote about it in greater detail (in the 174th edition of the newsletter), so I was interested to find out what has happened in the meantime.

Let's first check out some of the things that I said then. Here is one:

What I was never quite sure about was how a tool like this would make it into the league of the "big players" and compete. I did not quite see how to get freelance translators on-board and did not see much impact on larger players either. This all may just have changed with the latest release of XTM Cloud, a product that is geared toward translation agencies and freelance translators alike.

It's not really nice to start a conversation about a tool this way, but I'm afraid this strategy didn't work. XTM is doing better with larger customers (and even really large customers like Zynga), and naturally translators are using it if they are contracted by those clients, but it really hasn't won over the freelance market as paying customers, despite its not unattractive pricing of $43 for up to 60,000 words a month (less for fewer words and more if you translate more).

It also can't be the number of supported file formats, which includes virtually everything that other tools offer: all the Office, OpenOffice, HTML, XML, XLIFF, etc. formats as well as FrameMaker, InDesign, and text-based PDFs (converted to plain text).

It certainly can't be the support of operating systems: it's a browser-based tool, and while only Internet Explorer and Firefox are presently supported, you can use a tablet or a Windows, Mac, or Linux computer to work on the system.

Maybe it's the interface?

Here is what I said last time:

One of the most striking things about this product, which offers administration, bare-bones project management, and translation features, is the user manual, all 63 pages of it. That's it. And that's not because it's so badly written or written in shorthand or anything like that; it's because the product is very simple and straightforward. The interface is, well, Spartan and highly functional. It's not always completely intuitive (I was stymied with launching a project until I found a little "Start" button way down at the bottom of the page), but it's well organized once you get the hang of it.

Now, here there have been some changes. There are several manuals now, and the one for the administrator and project manager has indeed grown to 165 pages, reflecting a certain rise in complexity. In fact, as a PM you will want to spend half a day or so to learn the ropes and understand the process logic (that "Start" button is still where it was and once again I couldn't find it!). But the manual for the freelance translator is 10 pages long -- that's it, and that really is all you need to know about the tool as a translator.

The translation interface truly still is beautifully easy, even though some newly added features have been added to the mix (such as the display of up to 500 translation segments per browser page, the ability to join many different files into one, concurrent work on a file by several users, full history for every segment, more QA features, and a bunch of new machine translation engines).

No, it's not the interface that prevents translators from using the tool. Instead, it's the idea of an online interface that scares many of us, combined with the concept of our data being stored in the cloud, where it seems out of our control (admittedly, it might also be our clients who feel uneasy about this). And lastly, many of us are uneasy with an SaaS business model; instead of licensing a piece of software until we pay again for an update (while constantly fighting with computer and installation issues), we're being asked to pay a monthly fee (and have others worry about all the computer problems). I have to be honest: it has taken longer than I thought it would take to change these habits and notions of ours (after all, the title of the article about XTM was "This is the Dawning of the Age of...") -- but change we will, eventually.

Let's come back to the added level of complexity for the project manager and administrator: more features naturally require more settings, and somehow I find this especially true with browser-based tools, maybe because of the slight delay it takes to refresh a view or page. It's important to "get" the logic the developers applied, and it might not always match mine. I was surprised to have to look for certain things a number of times before I got the hang of it. (For instance, when I wanted to look at the machine translation options, it took me forever to eventually find them on the administrator level, only to realize afterward that, once activated, they apply only to newly created projects.) When I talked to the developers about the ease of use, they said that they're using the "Google principle: as easy as possible." Hmm. Maybe my logic is closer to Sergey's and Larry's . . .  in my dreams!

What's nice is that you can import data at every level, whether it's data about translators or clients or translation data in TMX or TBX (or other) formats. Also, the notification system for all project participants (translators, PMs, editors, etc.) is email-based and instantaneous. While you shouldn't expect a full-blown management system with reporting, accounting, and all the other bells and whistles, you will find tools to help you drive the productivity aspects of translation.

When I talked to the team at XTM, they prided themselves on one aspect more than any other: that "nobody understands XML like we do -- the Achilles heel of other systems." Of course, they're not just talking about translating XML files. They're also referring to the underlying data structure that the system is based on, in particular XLIFF, the format that all translation files are internally converted to (and externally: at any point you can export the XLIFF file, translate it offline in another tool, and reimport it into XTM). To put this insistence into perspective, you'll need to understand that Andrzej Zydroń, the main brain behind XTM, has been involved in the development of many of the important XML-based exchange standards, including TBX and XLIFF. So the XLIFF (I promise I will soon stop X-ing you out) that XTM produces is different from that of other tools because it doesn't use any of the extensions that make it hard to work with the XLIFF flavors of most other TEnTs.

I was surprised to hear Andrzej and CEO Bob Willans name the products they most often replace as SDL WorldServer for LSPs and (specifically in Germany) Across. This also provides a good idea of the size of the typical customer. If you are a translator who has recently worked in a browser-based tool, it's entirely possible that you have already worked in XTM. Because of its open API (application programming interface), it integrates itself nicely into other tools, such as the Easyling solution.

XTM has been and still is a translation technology trailblazer, and it would be lovely to see it find the recognition and success it deserves.

Here is another addition to my family of remarkable characters that I have been collecting over the years. This brings us back full circle to our introduction.

Hold your cursor over the character for a definition.

Why do people work with memoQ

"I'm sitting here at my desk doing some rather tedious translation work. Well, I wouldn't say that it is a delightful pastime on a Saturday afternoon - but memoQ makes it really bearable. It is such a pleasure to work with this translation environment."
Carola Striebel -- freelance translator

To learn more, attend free educational webinars and download the full-featured trial version from http://kilgray.com.

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 tentoeight.

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.

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. 

© 2013 International Writers' Group