TMX site

A brief history of TMX

TMX stands for Translation Memory eXchange. OSCAR (Open Standards for Container/Content Allowing Re-use) is the LISA Special Interest Group responsible for its definition. OSCAR members include translation tools developers, service providers and other interested parties (e.g. large translation clients). They came together over five years ago to specify a way in which translation memory data could be exchanged between tools and/or vendors with little or no loss of critical data in the process. OSCAR has recently voted TMX version 1.4 as an accepted standard.

What is TMX?

TMX is an XML format for the interchange of translation memory data. As such, it consists of elements (with attributes) that provide information about translation "segments". The size of a segment is not pre-defined and it will usually be a phrase, sentence or paragraph. For most tools using TMX, the default segment size is a sentence. Within each segment of TMX, there are optionally elements that provide information about the formatting contained in the segment (change of font, hyperlink etc.). TMX also provides for the definition of text "subflows" such as footnotes and index entries.

Example TMX snippet

Given the following HTML source and translation:

The TMX snippet might look as follows:

Is it easy to support?

Most commercially available translation tools already offer a text-based input/output format that can be used to transfer data to and from other instances of the same application. Making the necessary adjustments to use the TMX format as an alternative should be relatively straightforward as TMX is not a complex XML format and there are plenty of freely available XML parsers. Even if the tool provider is unwilling (or unable) to support the TMX format, it may be possible to develop complimentary applications that can do the necessary conversion between the input/output format of the tool and TMX.

Benefits

Exchange of memories

The most immediately obvious benefit of TMX is that it allows translation memory information to be exchanged between tools from different vendors (or, in some cases, different tools from the same vendor). This allows users the possibility of selecting the best tool for the job.

Example

Suppose that you have translated an old RTF help project using tool A. Your company has now moved to the newer HTML help format and you find that tool A is no longer the best tool for the job but instead you would like to use tool B as it provides better support for HTML. If both tools support the TMX standard, the translation memory data can be exported from tool A in TMX format and imported in to tool B with little or no loss of critical data. Note that TMX is not locked in to a particular file format (e.g. RTF, HTML), which facilitates this kind of process.

Choice

Providing a standard and encouraging its use means that you are free to change tools. Firstly, this ensures that you don't get locked in to a particular product with a proprietary translation memory format; especially if it is not commercially viable to use that tool any more. Secondly, you are essentially "future-proofed" as you will always be able to export your translation memories and switch to another tool if your current software vendor is no longer able to provide the necessary services.

Competition

Having a standard provides better competition. Without the VHS and DVD standards, I'm sure that we would find escalating costs and poorer features on our home entertainment systems. TMX helps to ensure that tools can compete on a level playing field. Without a proprietary database, products are left to compete in other areas such as:

• Features
• Functions
• Return on investment
• Support
• Price (!)

This ensures better tools and a more competitive cost for the end user.

Openness

With a clearly defined standard, other tools developers are given the opportunity to compliment existing translation functionality with new or proprietary features that can benefit the translation process.

Example

Suppose that you have a proprietary file format that is not currently supported by any of the available translation tools on the market. With TMX, you have the opportunity to either:

1. Create a translation tool to support this file format and then export the translations as a TMX file so that they can be cross-leveraged to other file formats.

2. Create a conversion tool to convert the proprietary file format to TMX format. You can then use a TMX-compliant tool to translate the text before exporting it out again in TMX format. Once you have the translated TMX file, you can use it to populate the translations in the original file format.

What you should do?

As an end user of translation technology, you should be looking to do the following:

• Ask your tool vendors about TMX

Ask them what kind of commitment they have made to conforming to TMX both today and as it continues to evolve.

• Get it in writing

Ask you vendor to provide you with their statement of intent to support the TMX standard and to help it evolve.

• Look for the TMX-Certified logo

Before you purchase any tool that creates or utilizes translation memory, ensure that it has been TMX certified. Only products that have been certified as conforming to the latest TMX specifications will be allowed to carry the TMX Logo.

• If it doesn't have the TMX logo, don't buy it

History has shown that your pocketbook sends the clearest message to tool developers.

• Build internal tools to conform to the TMX standard

A full description of the standard is available to you free of charge at http://www.lisa.org/tmx

David Pooley is a software architect at SDL International. He has been developing software for over twenty years and has been working on SDL products for the last five years.