Visit our company website at

Saturday, May 12, 2012

Why Computers Still Can’t Translate Languages Automatically

Why Computers Still Can’t Translate Languages Automatically - Slate, May 11, 2012
...The difficulty of knowing if a translation is good is not just a technical one: It’s fundamental. The only durable way to judge the faith of a translation is to decide if meaning was conveyed. If you have an algorithm that can make that judgment, you’ve solved a very hard problem indeed...

Wednesday, May 9, 2012

We all hate Trados, but...

As a supplier of translations to the translation industry, we are recently seeing a bewildering array of new translation packages being adopted by major LSPs.
Just to date this year, we have seen major companies moving to, and insisting on, one of the following: Wordfast, MemoQ, XTM, Catalyst, WorldServer.
Some companies have even gone further to completely reinvent the wheel by setting up their very own CAT tool with the view that their resource pool will be happy to learn their quirky and impractical system and translate online.
We all love to hate Trados, that is true. Many companies are trying to use Trados Studio 2009 and are not exactly delighted with the results, even after exhausting efforts to learn and implement its new features. Relying on a tool like Trados, that is created by and owned by a giant LSP that would like to take away all your clients, has always been irritating, but Trados has long been the industry standard and linguists have invested heavily in it.
At Admerix, some time ago we moved to using Trados 2007 in both WorkBench and TagEditor and all of our teams are equipped with this package, familiar and comfortable with its use, and the results are predictable and reliable. A quick perusal on any of the translator forums will testify that this is true, and while some translators do like other packages such as DejaVu they have to perform all sorts of file conversion tricks if they are going to serve more than a very few customers. Generally you can be very confident that if you ask a translator for a translation using Trados 2007 you will have no problem and the project will run smoothly.
What could it possibly be that is motivating the management of many LSPs to try to make expensive and time-consuming transitions away from Trados--especially after so many linguists have spent the time and money to be proficient in it?
The only answer that we can suggest is that such companies are desperately trying to retain a market share when new dominant technologies threaten to change the localization landscape. They inevitably turn away from good people (e.g. hire novice project managers and insist on the very cheapest translators) and towards some magical technological process fix.
The hope that “technology will save our company” has been dealt with on this blog in the past (in fact it was our very first post). Despite what management thinks, technology will not carve out a niche for both customers and a select pool of resources.
Our prediction is this process will spawn dissatisfied customers and make the vendor manager’s job practically impossible. Turnaround times will increase due to all kinds of technical glitches and slowdowns due to translators not being able to mesh with the new systems.
It also will limit the choice of translators who can be available for projects. This becomes a major problem when there is a need for specialist translator and editor teams. Also, and very importantly, scalability is limited when confined to such systems.