|
Most of you are
familiar with Microsoft's long and at times circuitous route toward terminology
sharing. When it started sharing most, if not all, of its user interface
translations in the mid-nineties (the files that were often incorrectly called
"The Microsoft Glossaries" were really translation memories), it was
widely, and rightly, welcomed as a very visionary step. While the files had to
be downloaded from an FTP site and were in a rather cumbersome comma-separated
(CSV) format, a good number of tools were offered that specifically or as an
added feature supported the particular Microsoft CSV format.
This was a visionary
step: Making these translation memories freely available ensured that all tools
running on the Windows platform would
use the same terminology in their translated versions, making it soo much
easier for users to switch between products. (When some tools -- I'm thinking
especially of SAP here -- decided definitely not to use the MS
terminology for political reasons, it only served as a sort of back-handed
confirmation of Microsoft's vision.)
After 12 years of
offering these databases to the general public, Microsoft suddenly withdrew
them and offered them only to (paid) subscribers of MSDN (and, later, also Microsoft TechNet).
The general public was first provided with a multilingual CSV glossary and a
few months down the road with the Microsoft Language Portal.
In its first
incarnation, the Language Portal offered access to terminology searches, style
guides in various languages, language-specific blogs (which were very
infrequently updated), and a sort of crowd-sourced site for the terminology of
some Microsoft products. At first, many viewed the site as a poor substitute
for the free and large databases, but it eventually became the standard for
Microsoft terminology. And, since the search queries were done through the
URLs, it was even possible to search with third-party tools like IntelliWebSearch (which unfortunately is not possible with the
otherwise very helpful TAUS engine).
Then last week, many
of you contacted me directly or sent cries for help to Twitter: The Language
Portal was gone! Alas, it was true -- but only for a few hours, hours that seemed
to last longer because the old URLs continued to be inoperative. For in their
place, under a new address, a completely new and handsome Portal had emerged.
I know it's very
easy to be critical of Microsoft (and Google and SDL and Apple and . . .), but
it's also important to give credit where credit is due. And credit is due right
here.
Aside from a new
look and an easier path to certain things (such as clear instructions for what
to do if you are interested in the whole set of TMs), many of the features have
stayed the same (terminology search, style guides, access to blogs). However, some
features have been updated and expanded (including the number of languages that
are currently covered or the commitment to more proactively publish new blog
postings), and some things are completely new, including the ability to
download bilingual files (English into other languages or other languages into
English) in TBX format.
Is that a big deal?
I think it is.
First of all, it's
super helpful to once again download data so that it can be integrated into
your own terminology resources/translation environment. And second, there is
the TBX element.
I realize that some
of you might ask what exactly TBX is. TBX, the TermBase eXchange standard, is
an XML standard that allows for the
interchange of terminology data, including detailed lexical information. The
adoption of TBX has gone very slowly, partly due to the fact that many felt it was
too complicated (for instance, see this article by Maxprogram's Rodolfo Raya, who has
developed his own competing and much simpler standard). Still, many tools have
now bought into it and are supporting it, including Across, Heartsome, Swordfish, XML-Intl, SDL MultiTerm, Wordfast Pro, Alchemy Publisher, and Star
TermStar. If you own one of those
tools, the TBX is easy to import (note that in the case of SDL MultiTerm you will first have to use the MultiTerm Convert program) -- especially since this is a TBX file
with a simple structure: just source, target, and definition.
But what about the
rest of us, those who don't have any of these TBX-supporting tools? XBench to
the rescue! XBench is a free
downloadable tool (I'm still waiting for the time when the makers of XBench start charging for it) that is
good for many different functions, including quality assurance of translation
files, lookup in glossaries, terminology databases, translation memories in
many, many different formats, and the import and export of such files. So it's
quite easy to import the TBX file and then export it into a TMX (Translation
Memory eXchange) or CSV file, which can then be processed by your tool of
choice (note that you might lose the Definition
field in this process).
So why is this so
big of Microsoft? Well, maybe "big" is the wrong term, but I join many
others in being thankful that Microsoft has reached out a hand to support this
important standard. What's also been very refreshing in my dealings with this
particular team at Microsoft is that bugs I've pointed out to them have typically
been fixed within a matter of hours -- not generally something I've been used
to seeing from a very large software vendor.
One more thing about
the Microsoft glossaries: They have also been integrated into the Evroterm termbase of Slovenia and the mighty EuroTermBank -- so if those are your preferred places to
search, you'll get the terminology that way.
Oh, and in case you
wonder which languages are supported (either in or out of English), here is a
list: Afrikaans, Albanian, Amharic, Arabic, Armenian, Assamese, Azeri (Latin),
Basque, Bengali (Bangladesh and India), Bosnian (Cyrillic and Latin),
Bulgarian, Catalan, Chinese (Simplified and Traditional), Croatian, Czech,
Danish, Dutch, Estonian, Filipino, Finnish, French, Galician, Georgian, German,
Greek, Gujarati, Hausa, Hebrew, Hindi, Hungarian, Icelandic, Igbo, Indonesian,
Inuktitut, Irish, isiXhosa, isiZulu, Italian, Japanese, Kannada, Kazakh, Khmer,
Kinyarwanda, Kiswahili (Kenya), Konkani, Korean, Kyrgyz, Lao, Latvian, Lithuanian,
Luxembourgish, Macedonian (FYROM), Malay (Brunei Darussalam and Malaysia),
Malayalam, Maltese, Maori, Mapudungun, Marathi, Nepali, Norwegian (Bokmal and
Nynorsk), Oriya, Pashto, Persian, Polish, Portuguese (Brazil and Portugal),
Punjabi, Quechua, Romanian, Romansch, Russian, Sanskrit, Serbian (Cyrillic and
Latin), Sesotho sa Leboa, Setswana, Sinhala, Slovak, Slovenian, Spanish,
Swedish, Tamil, Tatar, Telugu, Thai, Turkish, Ukrainian, Urdu, Uzbek (Latin),
Vietnamese, Welsh, Wolof, and Yoruba.
The amounts of
translated terms vary between 2,000 and 18,000 terms (you can find more
information on this in Jeromobot's Twitter stream).
|
