|
Close
to four years ago, I wrote an article in this journal about the release
of the open-source translation environment tool OpenTM2 (if you
have access to the archives you can read the full article in issue
169). Here's part of what I said:
LISA, the [now long-deceased]
Localization Industry Standards Association, announced the following
this past week:
(...) Teaming up with industry thought leaders,
LISA is announcing the OpenTM2 project.
Not only does OpenTM2 provide a public and open implementation
of translation workbench environment that serves as the reference
implementation of existing localization industry standards, such as
TMX, it also aims to provide standardized access to globalization
process management software. (...)
"There is a recognized and growing need for
standards in the localization industry. Despite our best intentions,
however, standards themselves can often be vague and open to multiple
interpretations. What is needed are reference implementations and
reference platforms that serve as concrete and unambiguous models in
support of the standard. It is our hope that OpenTM2 can become
a reference platform for the TMX standard," acknowledged Bill Sullivan,
IBM Globalization Executive. "Freelance translators are the backbone of
the localization industry. These translators have longed for free and
robust translation tools to increase their productivity. Our hope is
that by providing OpenTM2 for free we can enlist the aid of
this army of dedicated users to bring OpenTM2 even closer to
the TMX standard."
Doesn't sound too bad, huh? A free translation
environment tool to support the "freelance translators -- the backbone
of the localization industry!"
Many of you know, of course, that IBM's TM/2
was, along with Trados and Transit, among the first
translation environment tools. The tool was well liked by many of its
users, but in 2002 IBM decided to take it off the market and use it for
internal purposes only. There really wasn't too much weeping and
gnashing of teeth since Trados had already cornered the market
and TM/2 already felt a little outdated. Still, the ones who
liked it were naturally frustrated.
Now, eight years later, IBM has decided to throw
a pared-down version on the market as an open-source tool. In general
that's great, and it has received a lot of attention from various
sources on Twitter and blogs. So I spent some time looking at
the tool ....
After spending about an hour or two trying to
read through some of the documentation and make it work on Windows 7
(it is officially supported only through Windows XP) and really
not getting anywhere, I have essentially given up. The only conclusion
I can draw is that IBM and LISA have a slightly thwarted idea of the
"backbone of our industry" as a highly technical and very geeky (that
term again!) person who is interested in dealing with an outdated
interface and in translating only text and HTML files. Umm, not the
translator I know. . . .
Now
I would completely understand if you looked at the heading of this
article and the first few paragraphs and assumed that everything has
changed, that IBM has released an out-of-the box, fully functional
translation environment tool that everyone can download and use. I hate
to have to be the one to disappoint you. In fact, as of this writing,
the outdated website that hosts the
project dates back three or so years. Granted, some things have changed
since I wrote that first article, such as the additions of "additional
markup" (IBM-geek-speak for file filters) that now allow you to also
translate OpenOffice, JAVA properties, XHTML, XLIFF, and
generic XML files, but otherwise everything seems to be in a deep
slumber.
I
was really pleased when I learned from Gerhard Fetz from IBM's
globalization team last week that at least within IBM OpenTM2
is far from dormant. In fact, on April 23 IBM will start deploying OpenTM2
company-wide to replace IBM TM/2, the tool that is still used
(and cherished) within IBM. Gerhard explained that both tools (IBM
TM/2 and OpenTM2) are work horses - they're nothing you'll
win a beauty pageant with, but they can seamlessly process 13,000 files
consisting of 2.5 GB of data in one go (as he recently did in a
project). IBM's original plan to replace the existing technology with
something more modern was made virtually impossible by how deeply IBM
TM/2 was integrated into existing and well-tested workflows.
So
why the switch to the open-source version OpenTM2? For one
thing because the "giant pile of code" (or, according to Gerhard, "Riesenhaufen
Code") that was IBM TM/2 had all been sorted and ported
into C++ for the open-source version, which at its core is extremely
modular and still as reliable and automatable as the old pile and
ultimately extendable with its new plug-in architecture.
Still,
unless you are one of the many translators who actually works for IBM,
what does all this have to do with you? Well, I caught some renewed
enthusiasm as I talked to Gerhard.
For
one thing, I think I misunderstood the idea of the project at first --
or maybe it was just badly communicated -- but it really is not
primarily geared toward the individual translator, at least not as the
initial user. Instead, it's aimed at relatively tech-savvy translation
companies, enterprise companies with translation needs, and developers
who can implement and continue to develop this into an internal tool
that is truly able to move (data). Yes, there would have to be some
work invested into developing file filters (markup tables),
implementing it into the current workflow system, and building
extensions to any number of things -- including terminology components,
machine translation, shared data access, and quality assurance -- but
that might not sound so bad in comparison to spending $50,000 a year or
so in a support contract or licensing fees for a commercial tool.
According
to Gerhard, some of the items that will show up on a newly developed
website include a strengthened plug-in system, possibly more "markup
tables," a new "OpenTM2Scripter" designed to help build automation
scripts, as well as a tracking function in OpenTM2 that allows
the precise measurement of the time spent on translated or post-edited
segments.
It
also might be a good to consider that there are a number of freelance
developers who already have a very good handle on this system and can
start to work on needed components without any ramp-up time. Since this
is an open-source project, everyone is encouraged to contribute
personal developments, functional improvements, new plugins, education
material, a new user-interface, and new ideas in any way they see fit
-- something that can benefit you very much as well, as the example of
the translate5 developer in another article here shows.
As
far as the translators who would have to support a company's move to OpenTM2,
the biggest initial bonus is that the translation interface would be
free for them. Interestingly enough, quite a few of the translation
environment tools that have been around for a while still support the
good old IBM TM/2 format as well, so they could theoretically
even use those familiar environments.
I
know that there are other open-source tools, including the much-praised
OmegaT and the translation management system GlobalSight
(which are now actually integrated into
each other), but this is one other good resource, a little more
independent than GlobalSight and as reliable as the Big Blue
itself -- and as modern as you want to make it.
|
