Workforce Language Services

Conceptually this idea is great, but in practice I’m not sure what the value is. IBM has created a grant program to allow non-profits and schools access to their new web-based translation software. Bridging language barriers and improving education is great, but I’m not clear if this will accomplish the stated objective.

Imagine for a moment if the flood of emails you receive daily suddenly became an indecipherable series of fonts and symbols. Unfortunately, this scenario is a reality for many Spanish-speaking individuals with limited English skills. It’s a situation that is especially frustrating for parents when it comes to online correspondence with their child’s school.

IBM’s ¡TradúceloAhora! (TranslateNow!) grant program is offering an innovative solution. Using IBM’s WebSphere® Translation Server software, hundreds of schools and nonprofit organizations around the world are accessing ¡TradúceloAhora! to not only translate web sites from English to Spanish automatically, but also to translate emails bidirectionally (English-to-Spanish and Spanish-to-English).

And I am really not clear on what the value of this is over the many existing automatic translation software. And by reading in the FAQ section, I don’t get a better feeling for it.

How does the translation software used in ¡TradúceloAhora! compare to other available products?

It is as good or better than anything currently available. Automatic translation software overall is a “work in progress” and is far from perfect. However, those using the software can easily have a sense of the content on the web as well as access to the bidirectional email translations. With continuing feedback from users, IBM will continue making improvements that will increase the translation quality.

The translations contain incorrect or poor Spanish; why?

Language analysis and interpretation is a difficult task. Unfortunately, state-of-the-art translation software makes mistakes and doesn’t produce texts with the fluency that human translators have. The good news is that progress in human language technology is rapidly improving the quality of automatic translation. IBM is at work to continually improve the accuracy of automatic translation.

So, in their own materials they admit that their software is “as good” as what’s already in the marketplace (for free). And more importantly, the also state clearly that their translations produce translations that are full of mistakes. Maybe one day machines will be able to understand context, but until that day it will be very difficult for machine translations to replace true communication.