TechWhirl (TECHWR-L) is a resource for technical writing and technical communications professionals of all experience levels and in all industries to share their experiences and acquire information.
For two decades, technical communicators have turned to TechWhirl to ask and answer questions about the always-changing world of technical communications, such as tools, skills, career paths, methodologies, and emerging industries. The TechWhirl Archives and magazine, created for, by and about technical writers, offer a wealth of knowledge to everyone with an interest in any aspect of technical communications.
It depends a lot on your documents whether an off-the-shelf machine
translation package is total rubbish or maybe kind of useable. For
something you can show to the public without making a fool of yourself, you
need what one calls "professional translator". This is a human being which
is a good writer in his/her native language and has very good understanding
of the matter as well as in the source language.
The most useable tools are so called "translation memories". They are tools
for a professional translator. They are essentially a database of
previously translated phrases. When they encounter a similar phrase again,
they suggest the previously translated phrase, which may or may not be
appropriate.
The "machine translation" programs need - depending on the language - quite
a bit to a lot of post-processing. In order to "appreciate" unfiltered
machine translation, have a look at the Systran system provided by
Altavista. You may even look at the "translations" provided by that system
on the infamous Starr Report. From what I could judge, the French is kind
of understandable, with some hilarious terms, such as "costume de loi" for
"law suit". The German turns out as total gibberish, mainly due to the
completely different sentence structure of the German language.
There are some successful applications. The key for the success is a very
strict limitation of the source texts, such as using "controlled English",
which limits the allowable vocabulary to a few hundred words and phrases.
So, if your machine translation adventure should not end up in a total
marketing disaster, you would have to fulfill the following conditions:
1. Your source documents must be consistend and in a very simple language.
Idiomatic expressions, humour, cultural references etc. are an absolute
No-No.
2. Your sales offices must work with a professional translator, and they
also must very carefully check the results.
3. You and/or the professional translator must use translation memory
systems. General purpose translation software would need way more
post-processing.
One question remains, however. Why is it the sales organizations who are
responsible for the translation. With that, you (the writing department)
would no longer have the control. As the documentation is an integral part
of the product, you may get some difficulties in the quality control when
you do not have the control over these versions.
Add-on after reading the additional message: Forget it, if any of that
stuff goes out to customers. Particularly marketing material CAN NOT
REASONABLY BE MACHINE TRANSLATED without extensive post-processing.
Just my Zweiräppler.
Max Wyss
PRODOK Engineering AG
Technical documentation and translations, Electronic Publishing
CH-8906 Bonstetten, Switzerland
>Hi eveybody,
>
>My company's sales offices around Europe have asked me if there
>is a software package they can get that will translate text for them.
>One of my colleagues has an IBM package from somewhere that
>will do simple translations into German. He says he has only
>around 20% of text left to translate himself. Now that we are finally
>getting all our technical glossaries organised for all the different
>languages we work in, such a system would indeed be
>advantageous for the sales offices.
>
>Thanks,
>
>Jennifer
