The Translation Automation User Society (TAUS) is a group of like-minded organizations interested in advancing the state of the art for machine translation. Noting that organizations are being pressured to “translate ever more content in an increasing number of languages,” TAUS members (multinational companies, government agencies, and language industry suppliers) agreed to develop an industry body to host and share translation memories and terminology.

Over the next few months TAUS will explore how broad sharing of language data could increase translation automation through more leveraging and better training of MT systems. The group will develop a business plan by October 2007. It will propose an operating and financial model for the language co-op, assess the return on investment, outline the legal framework including intellectual property issues, and specify the required standards and technologies.

Why did TAUS decide on this path to a language asset co-op?

• It makes sense. For all our carping and caviling about TM Marketplace’s model for brokering translation memories, we do agree with the fundamental premise that there is mutual benefit in companies sharing their language assets. TAUS members see potential upside in accelerating time to global markets, quicker identification of terminology in different industry domains, and the resulting potential for lower translation costs and more satisfied information consumers.
• Self-interest. The TAUS decision is as much a defensive move against Google owning all the data on the planet as it is a good idea for the companies involved in any cooperative effort. If TAUS members don’t do it, someone else will — and Google is already sharing its corpora and MT data with users and with technology vendors such as Language Weaver. It could broaden that initiative to anyone in the same way it has made search universal against all manner of content.
• A reality check. The participants realized, as we have noted in the past, that these linguistic assets are already available on the web awaiting reverse-engineering by any commercial endeavor equipped with technology from a Language Weaver or MultiCorpora. Companies like Intel and Microsoft understand that they must take control of their language asset destiny before someone wrests it away from them. On the extreme end of this spectrum, witness Oracle’s complaint that archrival SAP has been downloading software and technical support information en masse from its customer support site.

Will the TAUS effort succeed? While reports from the group’s meeting indicate acclamation of the concept, the group still has a long way to go in reconciling the competing agendas and corporate interests — both intellectual property and financial — of member companies. It must also work its way through a thicket of legal, technical, and organizational issues. And it has to fashion an organizational structure that doesn’t bog down in parliamentary debates about payments, participation, new members, modifications, or new versions. But such an agreement is not impossible — if it were, you wouldn’t be reading this on an HTTP-enabled browser, the browser wouldn’t display Chinese characters, and your employer couldn’t pay you electronically through the Automated Clearing House (ACH) to sit in your chair absorbing knowledge. So it is possible for standards and industry bodies to make headway.

The bottom line is that this is a good idea — the challenge now is to make it happen by overcoming the myriad reasons why it can’t.