Using (or not) Google for Translation

From Ott09 Wiki
Jump to: navigation, search

Why to Google and why not to Google

An introduction: The just released translation tool kit was initially an internal translation tool developed by Google around 4 years ago – they have since put tons of resources and energy behind it, and then decided to take one step further to make it a public tool. They are very optimistic translations are 40% better with the tool kit, and quality will potentially improve the more people use it.

First we went on around to ask why everyone in the session was interested in Google Translate Toolkit + other translation tools by Google. Reactions were diverse, and interest ranged from some fascination with Google to suspicion, some curiosity about capabilities of the translation tools, some speculation on the future, will it support more languages (such as Bangla, Tajikistani)? What Google allow and doesn’t allow for translations needs; possible problems faced in the future; alternative to human translation; impact on the translation industry; potential conflicts; pros and cons.

It seems that with the arrival of Google Translation Toolkit and Google Wave there will be a new era for community of translators They say they are not charging for this service now, however, they might charge it at some point.

We all agree that the tool is good and has a great interface with lots of options – we can approach Google to find solutions for the problems we have and build on the top of them and ask them to make clear their usage license for translation memory, glossaries, end result translations, etc.

However, there were more question marks than solutions among the group.

The more TM is built, the easier volunteers’ life is. So what are the cons in helping Google to achieve better machine translation?

Google is in a very good position in solving the MT problem. Will they achieve it.

Professional translators’ have confidentially agreement with their clients, they will not trust the tool kit at this stage. It is a tool for amateur’s translations – Google is tapping in a long tail of individuals who can translate.


Approaching Google: What kind of organizations would be interested in offering their community expertise? And if our data is open anyway, would us matter to them?

Then the conversation then converged to another talk this morning: the open corpora repositories session.

To remember:

They depend on us as much as we depend on them. We should ask them to clarify licenses

Let's keep talking.