Box Journal subscribers save more on memoQ
you have your own memoQ translator pro license yet? memoQ offers 30%
off on all new translator pro licenses for Tool Box Journal subscribers!
our guide on the
purchase process and apply code CAM_TBJ30_PS in the memoQ
webshop. Offer expires at midnight November 15 in any time zone.
(How) Do You Use MT?
(or a similar) question was posted on a large pin-board at this year's
meeting of the American Translators Association (ATA) in Palm Springs
last month. I would like to share these with you uncommented (with only
one highlighted that was particularly clever: Like two porcupines
making love -- VERY CAREFULLY!) because these are all voices of
professional and successful translators (visiting the ATA ain't cheap),
which implicitly makes them important and reflective of a professional
in my CAT tool. Pre-translation when nothing found in TM. Needs very careful review.
quickly get the gist of a paragraph. As a quick dictionary when I
forget the target word in my native language again!
- For IT
messages and informational messages
curiosity! The robots are coming whether we like it or not. We should
know what they're good for and what not.
check genders (but we heard that's not 100% accurate...)
- For kicks
- To get
a good laugh
enhance the skill of human linguists to reason over language and
content and produce quality documentation of complex interactions
- To show
how it does NOT work and can never replace the high quality
professional work I do
in my CAT tool to assemble segments
two porcupines making love -- VERY CAREFULLY!
for any serious kind of translation work! For laughs -- to get the gist
of some Facebook posts and the likes only!
- To show
how algorithms can't make the mental leaps that humans can!
- As a
chainsaw, not a scalpel (i.e., only for a very rough job and not for
- PEMT of
popular TV subtitles -- at first I was incredulous that the domain
would process well with MT -- results were surprisingly good!
seldom as possible
- When I
am DTP-ing a translation in Adobe, and I need to understand the target
or match it up with the source
triage -- tells me at a glance what the topic of an article is. Will
sometimes line-edit MT output if nuance of source isn't important
don't! . . . yet
too!! Dinosaurs rule!!
- Talk to
when requested by the client
the client then on MT, CAT & difference
- . . .
and on human translation
un-user-friendly; terrible HMI
using a lever to move heavy rock -- you have to do the same amount of
work, but you can distribute it more intelligently. You don't have to
exert the brute cognitive force of creating words ex nihilo.
- I don't
-- is that wrong of me?
- Not at
all -- very smart of you!
it! I said "uncommented," but I can't help myself. Here are some
further thoughts in the next article:
treats, or the most desirable translation tools? No doubt, software for
translators is a win-win solution. Go ahead and benefit from the
exclusive offer for Tool Box Journal readers from Advanced
International Translations (AIT) at https://www.translation3000.com/aitpn/557-71.html
copies of each product available at 55% off. Get it or regret it!
on translation, not administration. Grab your copy now.
of all, I thought it was very interesting to see some outside-the-box
ideas in the comments (and the comments on the comments). I liked the
idea of using MT as a quick aid to know where to place text in a
document if you can't read the language, or for gisting, or as a quick
terminology resource. I also like the mentioning of different
strategies for MT for different projects and, again, I like the
porcupinian approach (you just learned a new word!).
is what I hope for -- and this ties in very well to a couple of other
discussions and presentations at the ATA: I hope that next year and the
year after and the year after that there will be changes in the answers
on similar pin-boards. There will surely always be the "dinosaur"-like
answers (and here I'm actually quoting a word rather than using a
descriptive term myself). And there will also always be the answers
that reflect the typical post-editing approach (editing one suggestion
from one MT engine) as the chosen way to using machine translation.
Percentages might shift a little toward the latter, but generally
speaking those two voices are going to stay unchanged. But I so long
for more outside-the-box ideas.
hope to see answers like using not one but several MT engines
simultaneously (or several suggestions from one MT engine); or using
auto-complete to just use fragments of MT suggestions; or using your
termbase or a glossary to automatically correct or flag MT suggestions;
or using your TM to give an MT match a reliability rating (or vice
versa); or using voice recognition to work alongside and together with
MT; or, or, or. (And here neither I nor anyone else knows what could
possibly follow those open-ended "or's.")
data generated by machine translation is a resource that can be used
with varying levels of success along with other resources; in fact, it
can be combined with other resources. Elsewhere in this Journal
I will talk about what we as translators have in common, but there are
so many things that make us different. In fact, one thing we do have in
common -- our creativity -- specifically makes us different in the ways
we approach our work. This work is so diverse, and that very diversity
requires different approaches for every translator, and maybe even
want us to be creative in the way we use our resources. Think about the
way we use and maintain our termbases. While it's true that overall we
probably don't use them enough ("too tedious and therefore too
expensive to create and maintain"), we have very individual approaches
to using them, even as to what kind of data we're entering, how we're
using the data, and what our hopes are for it. I think one reason is
that, while again it's sometimes underused, there won't be many
professional translators who will argue that terminology work is
useless or has a nefarious purpose. It's laughable to even think that
way. When I talk to translators about terminology maintenance, I
generally get only two responses. One is the slightly embarrassed
admission that "I'm not using it nearly enough, even though I should,"
and the other is "I love it and I can't imagine how I could work
there is no value in a terminology database per se. There is value in
what we make of it. And while this is not a perfect parallel to machine
translation engines, the "what we make of it" part of it is. Why aren't
we as adventurous in finding new and better ways to work with that
resource as with other resources? (Which, as in the case of termbases,
might mean that many don't use it.)
are some things that might help us: First we need to understand that
every situation is different. While it's great that large companies
train their MT engines and therefore have no problems with erratic
terminology in the MT output, that's not the case for the vast majority
of translators. There are many translators working for TransPerfect and
Lionbridge on the supplier side or Microsoft and GM on the buyer side,
but many, many more do not. This means that the experience of one kind
of translation does not necessarily "translate" to another set of
experiences. We need to really understand this in how we talk to each
other about machine translation -- and any technology. (Anne Goldstein,
the translator of Elena Ferrante and many other Italian authors, was
asked at the ATA whether she uses any kind of technology for her
translation, and before she answered "No" you could kind of see her
processing what the question even meant.)
there are real differences in language combinations. While neural MT
has leveled the playing field a little bit when it comes to the quality
of languages that are syntactically very different and were therefore a
difficult nut to crack for previous kinds of MT, it has for instance
not solved the problem of low-resource languages. (I regret not asking
for the language combination of the respondents to the ATA prompt.)
there is the kind of technology we use, both on the MT side and the
environment via which we access the MT (aka "translation environment
- Is the
MT adaptive or not?
- Am I
allowed to use certain kinds of machine translation?
- Does my
translation environment tool or my external plugin allow me to access
one or several suggestions and, if so, how?
- Do my
TM, my termbase, and my MT suggestions "talk" to each other?
- And so
on and so forth
the point on the technology we use: I realized in an ATA session on the
future of translation technology how certain choices by translation
environment tool providers play a pivotal role in how we can or cannot
use all kinds of things. SDL's Paul Filkin, for instance, mentioned
that while he felt really strongly about improving voice recognition in
Trados Studio for a long time (in particular the
interplay of Dragon and Studio), he no longer feels
this is so necessary after the recent improvements of machine
translation. I cannot disagree more! One of the reasons why I have been
very insistent on using the term "translation environment tool" is that
"environment" comes with the concept that all kinds of features are
available and can be used to the user's liking. It provides wide open
rooms that I can decorate as I like. And, yes, it's true that at this
point voice recognition and machine translation are unhappy bedfellows
(file that under "porcupinian"), but there's no reason why they have to
be, why there could not be better approaches for these two technologies
used in tandem. And it's unhelpful if a technology developer steers the
users' processes by essentially cutting one avenue off (especially if
the same developer is also selling access to MT services). Now the
truth is that every developer does this to some degree, and they
assuredly need to set priorities, but what we need to communicate then
is that we don't want to have our creativity and individuality stifled
by having our environments cut down.
that brings us to . . .
Translation with Memsource Instant QA
a small translation error can have a big impact. Memsource's new
instant QA checks, together with our robust batch QA tool, ensure that
every translation meets the highest quality standards.
The Conundrum of Having to Make Money With Machine Translation
heading doesn't refer to translators. Instead, I'm talking about
machine translation developers. I have a strong feeling that I will
receive a number of emails from said developers after this lands in
their inboxes, telling me that I'm oversimplifying matters greatly, but
here it is: I suspect that no one who is developing machine translation
is making real money from selling those products or services. The
exceptions are likely Google, Microsoft, Facebook, Amazon, Baidu,
Yandex, Naver, and (soon)
Apple -- but only because the machine translation products they produce
are just one part of a greater strategy. However, I doubt whether even
their MT departments would be profit centers if they looked only at
selling access to the data to translation provider and buyers.
same certainly is true for companies that specialize in machine
translation, with the likely exception of companies that get themselves
sold (such as when eBay acquired AppTek's, Amazon bought Safaba,
TransPerfect swallowed tauyou, or SDL picked up LanguageWeaver).
a really tough business, and the fact that now not only Microsoft but
also Amazon and Google offer customized versions of their engines
doesn't help the more specialized players.
really interesting example might be a machine translation provider that
goes back to the sixties: Systran. First of all, they are among those
who hopefully made a profit when they were sold to Korean CSLi, a move
that many thought was related to CSLi's client Samsung's interest in
machine translation, but we have not seen much of that yet, especially
when it comes to a Samsung-branded end user product. Systran itself has
come out with a number of new product offerings over time, including
the short-lived SYSTRANLinks, an offering for proxy translation
services (see elsewhere in this Journal for more information on
that). Now they are coming out with a new and, I think, potentially
interesting concept; whether it will be profitable for them remains to
talked to technical account manager Philip Staiger and Carolina Pinto
from Marketing at Systran's San Diego office about their hopes for the
product/service. It turns out that they might have announced a tad
early and/or we talked too early, though the timing might have been
good because I think they may have had some unrealistic expectations
about their product and its pricing. Here is what they know already:
They will be offering the possibility for you to create machine
translation engines from your translation memories that you can then
sell access to (or simply use yourself). They seemed less sure about
the legal underpinnings (minor things like data ownership, etc. ;-)).
The pricing also seems to be less than crystal clear. They will also be
selling API-based access (usable within translation environment tools)
to their generic machine translation engine in and out of 42 languages
(you can test it right here,
and you will likely find a similar translation quality to Google
Translate or DeepL). There is already API-based access, but it's on a
scale that is not affordable to freelance translators (between $700 and
$1200 a month), so the whole pricing scheme will be overhauled. When I
talked to Philip and Carolina they seemed to have unrealistic ideas
about what is affordable to translators, and I encouraged them to look
at some of their competitors to get a slightly better idea of what
would be reasonable.
we want everyone to make a living, right?
talk more about Systran's offering once it's released in its
style, your wording, your content. Utilize translations quickly and
reliably throughout your company.
machine translation powered by STAR MT
the short video for more information on STAR MT functionality and
years ago I wrote about a tool called Text
United that is developed by a Viennese-based company. I
recently had a chance to have a second look, prompted partly by a
number of changes in how the tool is being marketed and positioned.
is (mostly) a hybrid tool. In this case, hybrid means a locally
installed Windows-based desktop application that connects to
data (including translation files, glossary, and translation memory)
sitting in the cloud. The desktop application has an Outlook-like
Home screen, with the translation displayed in a
translation grid like so many other translation environment tools and
TM and terminology matches on the side. By default you'll need an
Internet connection to work in the tool, but it's possible to download
a local copy of your project and your resources (if you know you'll be
offline for a while) and continue to work offline. Once the connection
is restored, everything is synched and you can continue to work online.
are, however, additional alternative interfaces, including a
browser-based interface that can be used instead of the desktop app. A
key distinction is that you cannot work in the offline mode with this
browser-based interface and, if you are a project manager rather than a
translator, the browser-based interface does not give you access to
many PM-related functions.
a side note: I emceed a session at the ATA last week where the future
of translation technology was discussed. The people on the panel were
representatives from SDL, memoQ, and Wordfast, all companies with tools
that have been around for a long time. One of the questions we
discussed was whether the panelists felt they were missing the boat
when it comes to online translation interfaces. In that context (and to
strengthen my point), I asked the audience to indicate who among them
likes to work in a browser-based interface for their translation
environment. Maybe 10 or 12 raised their hand (out of a couple
hundred). I was really stunned (and obviously miserably failed to make
my point). After all, virtually all of the tools released in the last
five years or so are completely (or almost completely) browser-based,
and it's not hyperbolic to say that the majority of our other computer
activities happen in the browser. But I guess no one ever said that
translators are the most progressive bunch when it comes to our work
to Text United.
is also a third option for a translation interface, the so-called Overlay
Editor. This interface allows you to translate directly in a
website (so you can see context and sizing, etc.). This makes sense
because Text United not only supports a large range of file
formats (including MS Office -- notably Publisher! -- FrameMaker,
InDesign, all the tagged and software development
formats you can think of, and subtitle formats, but oddly not package
formats of competing translation environment tools) but also uses proxy
services to translate websites, ecommerce sites, and various other
content-managed sites. This method of translation takes a website in
its original language and produces it without actually getting into the
source of any of the translatable materials. Instead, the user browses
in a cloud-based version of the site without actually realizing it. The
cloud-based site continuously sends queries to the original,
untranslated website, which in response serves pages that go through
the cloud-based layer where they are translated on the fly and appear
in a different language. The real benefit to companies offering this
service is that they typically continue the hosting of the proxy sites
which assures an ongoing cash flow. The benefit to the translation
buyer is a potentially quick turnaround of their translation project
and an ongoing monitoring of any new content that needs to be
translated. Many of the larger LSPs offer this service with their
homegrown products as well as a number of largish and smaller
technology vendors, now including Text United.
this said, Text United's greatest hope might not lie so much in its
technological offering but in its new marketplace and infrastructure
plans. Freelance translators use the tool for free (their only payment
would be for the use of MT services that Text United resells, which
presently consists only of Google Translate but will include
other engines). The company also provides a marketplace of sorts where
translators can publish their profile and offer their services as well
as a quasi self-contained infrastructure that allows not only the use
of the technology but provides options for billing, reporting, and
project dashboards as well.
the six years since the last article, Text United has run a dual
business by offering technology and services. It has built up a base of
about 600 customers that buy translation through them and use their
technology to do that (the customers are primarily but not exclusively
in the DACH area). It's that growing network of buyers that's supposed
to tap into the freelance translators who are offering their services
so that they can directly contract with them.
Piorkowski, Text United's CEO, is convinced that the translation
industry is in the process of changing to where "translations are
integrated as part of the process platform." That process platform as a
tool as well as a manpower resource could (and he would say: should) be
I voiced some skepticism when we talked ("Do you know that virtually
every tool vendor has at some point tried to build a translation
marketplace and so far no one has succeeded?"), but we'll see what
happens. In a sense, they are in a similar position in relation to the
marketplace with an existing number of translation buyers as Across
with crossMarket (which has not taken off as much as was hoped)
and with a business infrastructure that reminds me of Smartcat.
way it's a reliable tool and even has some automated feature that
others don't offer as a standard feature, including an automated
terminology extract of your files' content that can serve as the base
glossary for the translation of the project. (Of course, you're free to
free access to technology is good, right?
Yahoo Groups Is Evolving!
was the message from Yahoo that many of you received in your inboxes.
That's complete nonsense, of course. Yahoo Groups, in my opinion the
only worthwhile part of Yahoo, is simply going away as we knew it.
That's noteworthy for translators in the context of this newsletter
because there are dozens of groups that particularly deal with
translation technology, and many of them have long been really
important backbones to the support for their respective tools and
technologies. (The support is/was typically of the peer-to-peer kind,
but some groups also have/had direct developer involvement.)
groups have already moved to groups.io or Google Groups, but I would
encourage you to look at your old and soon-to-be discarded Yahoo groups
and make sure you download all the content you might find valuable in
the Files section. You just might find some helpful macros or
tutorials there that may soon disappear.
hope that you are enjoying the '35
years of Trados' online festival!
translation-related resources and the chance to win great prizes.
spotlight: SDL Trados Studio Q&A webinar on November 28.
Follow Up On the Dictionary Exchange
want to give a shout-out to my tribe! We're all different (oh, so
different!), but I think it's safe to say that some things unite us. We
all love language. We all think of ourselves as relatively creative. We
don't mind working by ourselves. And: We like books!
the last couple of Tool Box Journals I promoted the Dictionary
Exchange at the ATA. And, really, Dictionary Exchange is
sort of a misnomer that doesn't really do justice to the generosity
that underlies the event. Although it's certainly possible to leave
some dictionaries and pick up others in exchange, in its essence it is
experienced translators and interpreters bringing their treasures for
younger, less experienced translators to equip themselves. (And when I
say "less experienced," I mean anything from translators who are just
starting out in the profession to translators who might be new to a
particular subject matter or maybe even a new language.)
knew last year's inaugural Dictionary Exchange would be a
success because people had mailed me large boxes full of dictionaries
to offer to their colleagues long before the ATA. Not so this year, so
I was a little apprehensive. Well, I shouldn't have been. It turned out
that conference participants had seen last year's success and just
brought the goods themselves. We ended up having hundreds of
dictionaries! (You can't actually see them all in the picture. There
are numerous boxes underneath the table of dictionaries that we
initially didn't have room for on the table, plus the first large batch
had already been taken.)
photo below shows the meager remainder (which ended up going to the
library system of California State University-Fullerton, thanks to
local ATA president-elect Madalena Sánchez Zampaulo, who had one
of the remaining dictionaries taken out of the box by an enthusiastic
Russian translator as Madalena was walking through the hotel lobby):
that the guy in the picture is not Madalena. I'm not typically in the
habit of publishing pictures of myself -- but after someone took a
selfie with me last week so he could post it on Twitter and then
afterward told me he didn't post it because I looked so
terrible in it, I feel like I have to make a point...)
two three reasons why I love the dictionary exchange so
much. First, because it is so low-tech. Clearly I'm not a techno-phobe
-- you wouldn't read this Journal if I were. But loving one
does not mean hating the other. Many dictionaries are available
digitally (either hard-drive-based or online), and if I have the choice
I prefer that method -- it's just a lot faster. But much specialized or
historical dictionary data is not as easily available, and for that we
have to use physical dictionaries. And (the second reason for my
enthusiasm) we do need to get to that specialized data because we're
actually serious and passionate about delivering high quality
translations -- exemplified by the oversize luggage fees many of you
had to pay upon leaving Palm Springs. Lastly, I love the fact that the Dictionary
Exchange is one of the great showcases for why and how we don't
compete with each other. It's because we don't have to. (What are the
chances that you have the same language combination plus specialty that
I have?) And even if we do: what a great opportunity to cooperate in
that niche field! When there is work for one translator pursuing
excellence, there'll be work for more.
assembling your dictionaries for next year's extravaganza in Boston at
ATA 61! (And if other associations want to do something similar for
their meetings: Feel encouraged!)
Last Word on the Tool Box Journal
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.
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.
you be interested in reprinting one of the articles in this journal for
promotional purposes, please contact me for information about pricing.
2019 International Writers' Group