ne of the areas at the heart of how successful a
Translation Environment Tool (TEnT) is for the translator, editor, and
proofreader is the quality of its word processing environment.
Early on, many TEnT makers chose the "easy way out" by using Word or
WordPerfect as their main word processing
environment. This seemed to have a number of advantages. Not only did
this provide access to the advanced word processing facilities that
came with Word and WordPerfect, but it also didn't hurt
their marketing message: "If you know how to use Word, you know how to
use our tool." This message really was a fallacy that badly back-fired
when users became upset that these tools were indeed quite complex and
challenging. Though it was nice to operate in a familiar environment
while writing or editing, the setup and maintenance of databases and
the use of all the many intricate features that most tools offer really
had very little to do with MS Word.
Wordfast is the only tool
left on the market that offers Word as its only word processing
Trados' Word editing environment
The first assumption—that Word does offer
advanced word processing features—was correct, though. So why is
there a move away from Word, even by the tools that are
automatically associated with Word such as Trados or Multitrans?
(In fact, to my knowledge Wordfast is the only tool left on the
market that offers Word as its only word processing
environment.) There are a number of reasons for this departure. One
obvious one is the reliance on a third-party tool that makes it
difficult to do your own independent planning for your tool. Another is
the need to require users to purchase Word if they want to use
your tool. And lastly, it is difficult to use Word with all the
many other non-Word-compliant formats that are to be translated
(DTP, tagged, other Office, database, software development, and
many other formats).
So, what features would ideally be part of a TEnT that
does not use Word as its translation editor?
Clearly, all the different language keyboards and Input
Method Editors (IME) need to work. While this is typically the case,
another kind of input via speech recognition does not always work as
easily. Even though speech recognition programs such as Dragon
NaturallySpeaking or Vista's internal tool work with most Windows
applications, including TEnTs, they are often less than optimal. For
instance, in some tools the beginning of a sentence is automatically
recognized and capitalized, while in others it is not; in some tools,
internal codes are harder to control with your voice than in others.
The development of a good translation editor needs to include testing
with the most common speech recognition programs.
We all know how important this is and we all know how
lacking most TEnT tools are in this regard! There are typically four
strategies that tool developers have employed so far:
- The use of a third-party spell-checker, particularly
the one by Wintertree software (www.wintertree-software.com). The
quality of this spell-checker is very language-dependent, and I have
yet to meet anyone in any language who is really satisfied with this
- Plug-in to Word to use its spell-checkers. Also a
relatively good solution—provided that you own Word and the
language pack for the respective languages (see below). To my knowledge
this does not yet work with Word 2007. The benefit of this feature is
that you can use the customized spell-checkers that you have in Word
anyway right away; the drawback is that it is usually a very slow
Choosing between the Wintertree and
the Office spell-checker in Trados TagEditor
- The use of plugged-in open-source spell-checkers that
are also used for OpenOffice (see
wiki.services.openoffice.org/wiki/Dictionaries). In many ways this is a
much better solution than the previous one, not only because of the
better quality of the spell-checkers but also because of the large
amount of covered languages.
- Browser-based tools such as Lingotek or Pootle
offer browser-built-in spell-checking (Firefox 2 by default and IE with
add-ons such as www.iespell.com or www.ie7pro.com).
Knowing how important good spell-checking is for
translators and editors, it is clearly important to find out how your
(prospective) tool handles this and how good it is for your language
(and at the same time for developers to use the broadest and best
solution available). Also, some languages have stand-alone third-party
spell-checkers that are often considered superior to any other product.
While this is no easy task, it would be wonderful if there were ways to
use these in the respective TEnT tools.
And lastly, only browser-based tools (see above) and across
offer the squiggly-underline on-the-fly spell-check that Word
or OpenOffice offer. It's true that some of us don't like this
feature, but others think it is absolutely essential so it would be
nice to have the option of using it.
Almost everything that was said about spell-checks could
be said about grammar checks, only that it is just never offered
outside the Word environment! I know that some folks REALLY
look down on grammar checks, but again, others use and like it. There
should be no reason why this cannot be offered in TEnT tools through a
link to Word or to some of the other third-party tools.
Grammar checker in Word
AutoText and AutoCorrect
AutoText (the ability to expand a token with a keyboard
shortcut) and AutoCorrect (the feature to automatically correct common
typos and other customizable short forms) are part of most word
processing tools and should also be part of any TEnT tool. Typically
one or the other is done by most tools, but rarely both. It should also
be possible to import language-specific lists from other tools.
Déjà Vu is one of
the only tools that offers AutoText and AutoCorrect
and imports entries from MS Word
This is huge and it's really a no-brainer that it should
be part of any kind of TEnT tool. Yet at this point, not a single one
provides it outside of the Word environment. There are
workarounds (some tools track whether a different user has touched a
cell, while others have third-party tools available for a comparison),
but an elegant integrated solution needs to be part of any word
processing environment for any tool. The very nature of the job makes
this is a necessity, because the translation process typically involves
editing and proofreading passes during which the documents are
exchanged back and forth between translators, editors, and
Though it uses a different process than Word, Lingotek
also stores information about the different stages and allows you to
compare them in any combination at a later point (even through its
exported XLIFF format).
Comparing files in the third-party
application ApSIC Comparator
Comments have fortunately become a rather common feature
within TEnT tools, and the few that do not offer this feature should
definitely do this. It's just very handy to add some comments as you're
working when you are in doubt as a translator or editor (or you want to
express your appreciation for a colleague's good work . . .).
Comments in MemoQ
Inline codes are the markers that are used to "remember"
formatting or other kinds of tags within sentences in a TEnT
environment. Every tool needs to deal with those in some way or other.
What's important is that placing these codes not interrupt the workflow
of translating. Anything that can be achieved only through the mouse is
not user-friendly and should be banned. It is important to have
(customizable!) keyboard shortcuts for this task.
This seems like a small thing, but most of us know it is
very frustrating to have to enter the smart/curly quotes of our
particular language with the help of some kind of fancy work-around.
The tool should do it for you.
What-you-see-is-what-you-get—the ability to see
the translation in its context and layout—is as old a topic as
TEnTs, and different users simply have different preferences. What has
become common practice for a good number of tools is not to offer a
complete WYSIWYG environment while you translate (because this could be
distracting and is in some cases just not possible, such as when
dealing with text that is embedded within certain code). Instead, the
user is given the chance to switch on the fly to a WYSIWYG environment
(as much as that is possible) to verify the placement and context of
Translation and Preview view in Idiom
WorldServer Desktop Workbench
In an editing environment it's helpful to see these to
avoid or make sure of double-spaces (within a sentence or after a
period), to see the difference between a breaking and a non-breaking
space (one of the few ANSI combinations everyone should know:
Alt+0160), or to verify whether a tab is used instead of spaces. Transit
has a very fancy and user-definable implementation of this and so does SDLX.
Trados TagEditor shows normal spaces vs.
non-breaking spaces and so does across by default. Most other
TEnTs don't do that, though.
Setting non-printing characters in Star
I can think of a number of other things that could be
listed here, but if the items above were to become a minimal
best-practice checklist for tool developers, we all would be well