|
My
long-time colleague and friend Juliet Macan recently wrote me this
letter. I found it remarkable enough to share with you because it
describes the world of translation in a way that only few of us have
actually seen. Here you go (and thank you, Juliet!):
Your
book "Translation
Matters" really triggered some amazing memories. Quite literally
as someone who experienced the industrial revolution of translation, I
wanted to share my experience as a translator and translation
technology enthusiast living in Italy since 1973 when this marvellous
adventure all began for me. Here is my viewpoint of how things
occurred, at least in Southern Europe:
Although
I never had any formal (academic) training in translation, I studied
Sociology and Psychology at university. I found the great variety of
texts I was asked to translate as a junior translator in an Italian
agency an endless challenge and delight. Curiosity was the name of the
game! Delving into the meaning of elaborate Italian texts which sang
beautifully in the original but translated literally would have sounded
plain silly to an English ear. Having to extract and mould the message
to another culture without losing possible inferences. I loved it. My
fascination with words harks back to my childhood, where I soon
realised that some of the household words used by my Irish immigrant
family were quite different from those used in my English school. This
together with my Catholic Latin upbringing, passion for reading, and
love of poetry and singing made me particularly sensitive to language
and meaning.
But
it was less fun having to retype one or more pages on my manual
Olivetti typewriter when a lot of changes were required. So I was
thrilled, in the early eighties, when I first got my hands on an
electronic word processor (Rank Xerox), so I could type up and save a
certain amount of text and modify and reprint it as necessary. This was
just a short jump to the PC. Unfortunately, my first computer, another
Rank Xerox, did not go so well: after losing a translation of 30 pages,
I sent it back but, undeterred, opted for my first IBM clone. Here I
went for Microsoft Word for DOS, and my computer
adventures began!
Certainly,
there were hitches in the beginning, but with a group of fellow
translators, I learnt to master MS Word for DOS, to format the
layout of a table in pre-WYSIWYG days. Lots of adventures in this new
digital world. Also, instead of handwritten notes in my dictionaries,
or pieces of paper with terminology suggestions, I was able to create
"real" glossaries in my PC. But, just like all my reference materials,
the original texts were still in paper form and the translations had to
be delivered by post or in person. One nasty drawback that I
experienced in my early computer (and mouse) days was "tennis
elbow"--extremely painful--due to the incorrect position of my keyboard
and mouse. Once I learnt the importance of ergonomics, problem
resolved. Strangely enough, 15 years on when Dragon
NaturallySpeaking was available, it was no use to me, as by that
time I needed my daily exercise on the keyboard to keep arthritic
fingers supple!
Then
the BBS arrived in the early nineties, which meant that equipped with a
modem, I could deliver my translations in electronic format without
leaving the house! Although I did have to install a second telephone
line so that I could explain to the agency, or client, how to receive
my file. This was rapidly followed by internet: finally, I could send
and receive documents in electronic format. This new era changed lots
of things, and required increasing technological competence, so I
joined a cooperative of translators and interpreters so that I could
contribute to, and benefit from, the pooling of knowledge and
investment with professional colleagues. The arrival of MS Word for
Windows introduced new changes for the better along with
challenges--the RTF that lies below a DOC still causes trouble today
for those who refuse to move on to DOCX. Other challenges were DTP
formats such as Adobe PageMaker and FrameMaker and Interleaf,
etc., as technical writers began to provide their documents in these
programmes. Either you installed these programmes and learnt how to use
them, or you were out of a job! But it was often difficult to focus on
the translation in hand when you were struggling with unfamiliar
programmes as well as layout problems, as clearly many of the tech
writers were focussed on the "appearance" of their documents and
unaware of the "text flow" nightmares they were creating. Somehow I
feel this was excusable in the nineties, but I really think it is
unforgivable today, especially in MS Word files where tables
can be excruciating.
Simultaneously,
the arrival of terminology and translation memory software created a
revolution. The potential benefits were very clear, but there were many
new problems, not least choosing the "right" tool and configuring file
filters, not to mention compatibility. IBM, Star, Trados and Atril all
offered translation tools!! What a tumultuous time! I admit, I just
jumped in the deep-end, tried a bit of everything, and gradually
refined my choices over the years. Because of this passion and
curiosity to find the best solution, at the end of the nineties I was
appointed the translation technology manager in the cooperative,
responsible for testing new systems and utilities, overseeing the
fine-tuning, and training translators and reviewers in the use of these
tools. This aspect had become even more urgent following the
introduction of the new European Union laws in 1995 that obliged
companies to provide all their documentation in the language of the
European countries where they sold their products, thus no longer just
English, French, and German, but Finnish, Danish, Dutch, etc. Another
revolution that multiplied the demands for translation of technical
documentation and required the training of new translators.
Then,
in 2000, I had to find a replacement for IBM Translation Manager
when IBM decided to suspend support. Although we used Trados,
and TagEditor, for quite a lot of work, I was one of the few
who first used and appreciated Workspace as it offered the
possibility of creating multilingual projects and combining the various
parts of Trados into a single tool. I was getting exasperated
by the total lack of support offered by Trados. Then there was
the question of the freelance translators who did not have a licence,
or those who swore they worked with it but who used Wordfast,
which meant serious segmentation and compatibility issues when they
delivered their files. Exactly at this time, SDL put its in-house
translation software on the market. SDLX was a great solution.
It made it possible to create multilingual projects and provide
translators with a free translation environment, with everything set up
and connected to aid them in their work. The translation layout was
much simpler and more pleasant than IBM Translation Manager, Word+Workbench/TagEditor,
and Transit, and quite similar to Déjà Vu.
Support was also excellent: if they couldn't fix a problem immediately,
they gave you sympathy and a workaround. They appreciated feedback!
Furthermore, it was easier to convince translators to try using CAT and
appreciate the advantages; all they had to do was open the packages and
translate.
In
the meantime, not only did Microsoft update its basic Office
tools to the XML format, another improvement, but it was also the
advent of various so-called "QA tools," another source of great
excitement and exploration. QA Distiller, Dog Error Spy,
ApSIC XBench, etc. offered new possibilities for
helping translators to avoid mistakes and improve the accuracy of their
translations. Each of these had to be assessed, to establish how they
could be used to best advantage, and instructions prepared for
translators and reviewers.
Then
there was the new enlargement of the European Union in 2004 and the
requirements for translations in the "new" European languages, from
Hungarian to Czech, from Polish to Maltese (for the latter it took
Microsoft from May to September to include the four special characters
unique to this language). Even more translators to introduce to
translation memory, and new fonts to be designed for FrameMaker!
And
then the breaking news on June 20, 2005: "SDL acquires Trados." I knew
immediately what this meant, and that I would be obliged to look for
another tool!
A
lot of new translation memory tools had appeared, such as MultiCorpora,
OmegaT, Across and Heartsome. But
none of these felt quite right. Then I tested memoQ, and my
first impression was that they had clearly learnt a lot from all the
tools that preceded them and appeared to instinctively know what was
efficient and comfortable for a translator. In addition, they also
appreciated feedback and were prompt in implementing PM functions.
Subsequently, I was interested to note the similarities with the new SDL
Trados Studio, although I think that Studio was based
heavily on the experience gained with SDLX, with bits of Trados
plugged in. They certainly had, and still do, keep a close watch on
what memoQ does. This has led to benefits for all over the last
decade, as both tools have introduced features that they considered
useful in the other tool. The only oversight, which I consider
appalling, on the part of SDL is their failure to recognise the
incredible possibilities offered by LiveDocs.
Now
to get back to "Translation Matters": one of the first things that
really touched me was your reference (in 2010) to the cloud (pages
38-40) and your doubts. Even though the cloud has been around for over
a decade in certain tools and lies at the base of many of the newer
tools that have appeared since 2010, I still feel uncertain about
various aspects. First of all, although I personally have a fantastic
fibre connection to the internet from home, I am acutely aware that not
everyone is as lucky. Secondly, when you are "on-the-road," connections
can be very unreliable/unpredictable, and so if your main tool, and all
the material you are working on, is in the cloud, even in 2019, this is
not a good thing. Indeed, I had a discussion about this with Michael
Farrell in Forlì, as he was looking for the ideal translation
environment tool with everything in the cloud. I tried to convince him
that this was probably not a very good idea. And it is not even just a
question of lock-in, as you mention later on, but the fact that even
though you may be able to access, with permission, a cloud from a
different tool to do a job, at the end of the translation, you are left
with nothing. No trace for your own reference, no idea of how the
translation has been used/modified. As a translator I find this more
than a little alienating. Furthermore, it is not necessary. You can
work in the cloud, but still have the possibility to work locally.
I
was delighted with your mention of Terence Lewis, an amazing guy whom
we have had the fortune to hear speak twice at the ASLIB/ASLING
conference in London. He really fascinates me with his determination to
find a solution for whatever problem crops up. On the other hand, I am
still in two minds about crowdsourcing. While I fully agree that it is
a marvellous opportunity in certain circumstances, for instance
Wikipedia, I was very disturbed when memoQ introduced several
years ago the online professional workflow by which a translator starts
a translation and then, to save time, almost simultaneously, a reviewer
can enter and start to check and modify the translation. Maybe I am too
old for this but I cannot escape the feeling that when one is doing a
translation, often you need to get into a text before you can be sure
of the correct translation, then you need to go back and review it
yourself, before passing it on to someone else. But if somebody else is
pressing hard on your heels and changing your translations before you
have had time to reflect -- I am not sure what the result will be. I
also notice today an unfortunate drop in the sense of responsibility on
the part of translators for their work, probably linked to project and
payment structures. This is dangerous and very inefficient.
As
regards your Trados history, I think we need to get some of the
facts straight. The company Trados was founded as an LSP in 1984 and
therefore could boast a history of 25 years in 2009, but the software Trados
Workbench for DOS first appeared in 1992 followed by Trados for
Windows in 1994. So at the most, we could say that in 2009 Trados
Workbench was 17. This is something that SDL, which launched its
own project-based translation tool SDLX in 2001, plays a lot
on; indeed, Studio (a combination of the two) is only 10 years
old this year!
I
very much appreciated your Relationships section, especially now when
the pressure is on to automate everything from uploading content by
customers, automatic counting and assignment, and automatic delivery
and billing, although strangely prompt payment seems to be lagging
behind! Clients rarely know much about the intricacies of languages,
let alone translations or translation technology, and neither do their
staff or content authors. To pretend that all this can be done by a
magic tool, which just as soon as NMT has been perfected will do away
with humans altogether, is absurd and yet.... I agree that qualified
PMs are important to facilitate the whole process; they are NECESSARY
to oil the wheels. Indeed, when I was still working in an LSP four
years ago, I was delighted by the possibility offered by the memoQ
project-communications function, a kind of Skype conversation for all
those involved in a project that allows translators to discuss
terminology or other aspects of the content with each other, and which
can be monitored by the PM who can step in when necessary, or refer
questions to the client. It was much appreciated by the translators and
the PMs. But even at that time, the pressure was on to automate and
eliminate human interaction.
The
message needs to be that communication, in one let alone multiple
languages, requires the comprehension and mediation of highly qualified
humans. Machines can follow patterns, but they do not understand.
Unfortunately, often the problem lies with the source content. Despite
the valiant efforts on the part of many authoring and content advisors
over the years, source content optimisation is lagging sadly behind. A
typical example of this can be found in the widespread use of
Euro-English as a pivot between the various EU languages. This,
combined with the pressure to use NMT and PE, displays complete
incomprehension of the issues. Yet another argument in favour of the
dream of Interactive Translation mentioned by Félix do Carmo
(see edition 299 of the Tool Box Journal) to support expert,
specialist linguists in the identification and solution of
communication problems that cannot be dealt with by AI or someone just
fluent in the target language. We were lucky enough to have
Félix as a speaker at the Translating and the Computer
conference in London in 2016 and I found his words very inspiring.
You
can find Juliet on LinkedIn right
here.
|
