TMX: Maximizing the Return on your Translation Memory Investments

Standards are the foundation upon which technology innovation thrives. This is because they allow more advanced technology to be developed by assembling standard components. The new, more advanced technology then requires new standards. This is how an information industry builds its conceptual edifice, layer by layer. It is a sign of maturity in any industry when standards start to appear, and our industry is no different.

-Pierre Cadieux, PEOPLE, Process and Technology

Why Now Is the Time for Standards in Localization

High-tech product generation cycles have shortened dramatically with leading hardware manufacturers such as Dell and HP producing hundreds of variations of their machines every few months, all with localized documentation and sales support required as part of a total solution for customers. The same applies to software, where the production cycles are now counted in months. Not to mention web sites, where new content can appear every few hours or days, all in a number of languages that must be updated simultaneously.

And the definition of content has changed over the years. What is it today? Unstructured and growing by leaps and bounds in the form of streaming content of bits and pieces, audio, video, images, etc. The software industry, in particular, is moving to a “sustained release” model where there is just a constant stream of content. The good news is, that whatever format it’s in right now, it’s all just information, waiting to move onto the next medium, the next culture, the next language… all with our help as enablers.

Translation memories have become critical to the business case for localization.

What allows companies to keep up with this constant barrage of content in its many formats during the localization process? Translation memories. They have become critical to the business case for localization because they allow organizations to store their linguistic Intellectual Property (IP) to be re-used and re-purposed across multiple platforms (including traditional publishing environments, content management systems, etc.) and across multiple languages (HTML, XML, etc.).

TMX: The Only Standard That Creates an Open Environment

TMX allows for the exchange of translation memory data, with little or no loss of critical data.

The only way to make this possible is through the adoption of standards to ensure that translation memories (TMs) can be shared across all of these environments, without being locked into any one company’s proprietary format. There is one standard that creates this open environment, and it is starting to make a real difference for customers, language service providers and technology vendors alike: Translation Memory eXchange (TMX), the vendor-neutral open standard for storing and exchanging translation memories created by computer-assisted translation (CAT) systems and other tools. Editor’s Note: TMX was the first publicly defined and available standard based on XML (itself not finalized when work on TMX began!). For a brief history of TMX, read TMX – A Standard Ahead of Its Time, by Arle Lommel.

TMX is an XML format that allows for the exchange of translation memory data, with little or no loss of critical data, between tools and between language service providers, across platforms and across products. It is clearly documented and incorporates a verifiable, third-party testing scheme. The standard is sanctioned by the Localization Industry Standards Association (LISA) and was created and is maintained by its Special Interest Group, OSCAR (Open Standards for Container/Content Allowing Re-use).

If you have heard of TMX, but don’t really know what it is or why it’s important, read The Importance of TMX, by David Pooley, Software Architect at SDL International. Details on the current version of TMX are available in this article by Yves Savourel, author of XML Internationalization and Localization and Localization Solutions Architect at ENLASO. Additional information is available in the LISA/OSCAR Translation Memory Survey.

Over time (and much sooner rather than later) more and more organizations will migrate to single-source authoring and multi-channel publishing, all built on XML. Since TMX is XML-based, it is (and will be) the vehicle for seamlessly exchanging TMs between the authoring, content management and translation environments.

Savitha Varadan, Manager for Global Web Content at PeopleSoft describes how PeopleSoft has integrated a CM system with SDL products, based on Certified TMX, to centrally manage the publication of the company’s web content in 10 languages across 24 websites:

On a scale of 1 to 10, with 10 being the highest, the importance of TMX was about a 7. We need the capability to be there in case we decide to share translation memories internally within PeopleSoft in the future. The actual implementation and performance of the technology components from SDL have surpassed our expectations, and we are finding that it is truly scalable. It works and performs according to spec. This happens so rarely with software, that one is always a bit shocked when it does.

Why Has TMX Become Critical to Protecting Intellectual Property?

In today’s world, words represent a significant IP investment for global businesses.

In today’s world, words (translated into multiple languages) represent a significant IP investment for global businesses.

TMX is the solution to the commercial problem of protecting this investment. TMs cost money to generate and to maintain and should be the property of the organization that owns the source and target content. Through the use of TMs, words come to be treated as a capital investment since a certain value is placed on each word as it is first created, then translated, then stored and then reused again and again. It follows that a certain value must be extracted from each word over a period of time to recoup the investment. Therefore, it is only logical that the words stored in TMs should be available and reusable in whatever format and through whatever tool the customer may choose.

TMX has become the key that unlocks the power of this investment in words and protects the investment over time. Without TMX, the words remain locked away and trapped, reliant on proprietary technology. Through TMX, users can select from many different tools and can change tools whenever they want, allowing them to always be able to use the best tool for the current job.

Chris Stockdill, Documentation Manager for Surfware (one of the world’s top ten fastest-growing computer-aided design manufacturers) explains how TMX protects his company’s IP assets:

The more languages we support, the more important it is that we leverage the latest technology to optimize the localization process, SDLX incorporates many powerful features for localizing software, including tag protection in help files and the ability to hide non-translatable code. Its easy-to-use interface improves productivity and its TMX certified translation memory protects our investment in translation assets.

Standards such as TMX are extremely important because they ‘future-proof’ the investments of customers, language service providers and tools vendors.

Customers, of course, are extremely pleased that they can lower the total cost of ownership for their linguistic software assets. As Mika Pehkonen, Localization Manager at F-Secure Corporation, explains:

We ensure the technical compatibility and cross- project reusability of our most important asset, our translation memory base, by having our partners deliver in a well defined format that is universal – TMX. In a corporate environment, where translation memory is the key to consistency, cost savings and return on investment, SDLX provides us with the benefits of TMX Certified translation memories, added security, and peace of mind.

Language Service Providers

Language service providers can now concentrate on value-added products and services, rather than losing time by working with TMs in multiple formats.

Aaron Schliem, Director of Operations at Glyph Language Services, explains the value of TMX for his company:

We had a long, hard look at the future of translation technology and it seemed that the only things the industry was talking about were industry standards such as TMX and the complete solution in SDL Localization Suite. With our rapidly growing customer base, SDL Localization Suite provides us with all of the required tools to effectively handle any project our new customers send us. Along with the guarantee of a TMX Certified product range, we’ve already been able to take on more client work and secure key wins with new customers.

Tools Vendors

Tools vendors are freed to focus on innovation and to compete on features and functionality since TMX provides a level playing field.
Tools vendors no longer have to waste time supporting multiple TM formats.

SDL has been instrumental in driving the development and adoption of TMX. Terry Lawlor, Vice President of Worldwide Marketing at SDL believes that supporting open standards such as TMX will grow the overall market, as well as free up users to choose the best tool for the task at hand.

When an industry adopts standards, innovation thrives – just look at how the Internet has exploded because of standards like TCP/IP, HTTP and HTML. When everybody is certified to TMX, less resource handling for multiple proprietary formats will be required, so we will be able to devote more resources to developing the products and features our customers need. At SDL, we are determined to lead the way with industry standards.

With time being saved by all parties involved, TMX contributes to taking cost out of the localization process. It also facilitates localization outsourcing since TMs can be seamlessly shared between all groups involved. All these factors will help to grow the entire localization market.

How to Ensure That Products Are TMX Certified

Only products that have passed the third-party certification testing process may claim to be TMX Certified and thus entitled to display the TMX Certified logo (Release 1.4a or 1.4b). Currently, SDLX, SDL International’s CAT tool, is fully certified, and Déjà Vu from ATRIL is about to be certified. Other tools are starting to go through the certification process.

The best way to ensure that products are TMX Certified is to vote with one’s checkbook. When customers require that products be built on open standards such as TMX, this forces vendors to listen. As end-users of translation technology, customers should do the following:

Look for the TMX Certified logo before purchasing any product. Only products that have been certified as conforming to the latest TMX specifications are allowed to carry the TMX Logo.
If a product doesn't have the TMX logo, customers should not buy it. It bears repeating: closed pocketbooks send the clearest message to tools developers.
Move translation memories into a certifiable TMX format (1.4a or 1.4b), so that they are no longer in a proprietary format and can be accessed by a wide variety of tools and utilities. In this way, they will be a much more valuable asset to the business. A new utility, TRADOS® 2 TMX, has just become available to do just this. Build internal tools that are based on the TMX standard so that you only have to write them once. This will allow them to work with translation assets from multiple vendors and many different sources.

With customer demand for vendor independence growing and an increasing number of tools spanning authoring, content management, CAT, QA and review now being used during the localization process, TMX should be a critical component of every customers’ localization strategy. The business case is compelling for customers:

Invest the time in adopting the TMX standard, and you will enjoy a handsome return – the value of your translation assets will increase, and your localization costs will decrease.

Just as important, you will be making a positive contribution to the overall health of the localization industry!

Editor’s Note: Because TMX and other standards can serve so many different purposes, OSCAR is holding a special session on June 22 during the LISA Global Strategies 2004 Summit in California on the future of open standards in the industry. This is your chance to influence the future of standards, so plan to attend and participate in the “visioning exercise” that OSCAR’s chair, Gérard Cattin des Bois, has proposed.

If you’re interested in hearing what SDL has planned for the future, don’t miss Terry Lawlor, Vice President of Worldwide Marketing for SDL International, when he speaks on Software, Services and Knowledge-based Translation – Localization Solutions for Global Businesses on June 23 at the same event.

Reprinted by permission from the Globalization Insider,
17 June 2004, Volume XIII, Issue 2.3.
Copyright the Localization Industry Standards Association
(Globalization Insider: www.localization.org, LISA: www.lisa.org)
and S.M.P. Marketing Sarl (SMP) 2004