Recent Posts View All

Blog - The loneliness of the long distance standards committee

The loneliness of the long distance standards committee

posted 2007-02-16 11:00:21
by Peter Reynolds

Sometime soon, at least I hope soon, XLIFF will be published as a standard for localization file interchange by OASIS. After six years of work the finishing post is in sight, specifications have gone through reviews by our peers, comments have been made and responded to and we now have something that helps solve problems faced by those who translate their electronic data. Although it’s not really a finishing post. When we have got past the stage of releasing XLIFF 1.2 as an OASIS standard we are starting to work on the next version. Now seems like a good time to look back on this committee.

Work on XLIFF started when a number of companies based in Dublin, Ireland decided they would like to work together on a project. The lets work together on some project decision was made at executive level and it was left open as to what that project was. Paul Quigley, who at the time worked for Oracle, suggested that one problem was too many file formats, too much complexity and it would be great if we could just simplify this with an interchange file format. The more we thought about this, the more we realized how big a problem this was. During an early discussion on file formats, companies like IBM, Novell, Oracle and Sun mentioned large numbers of different file formats which they need to localize. The attraction of simplifying this to have a single file format for localization looked like the sort of holy grail we should aim for.

We realized that we were not the first to try and tackle this issue. Yves Savourel had worked on a XML format called Open Tag which seemed to us like the father of XLIFF. We got in touch with Yves to ask him a few questions. He liked what we were doing and joined our committee.

At the early stage we had two names. The first was data Definitions and the second was LIFF (Localization Interchange File Format). Data Definitions was a very early name which only lasted because we had called our yahoo group by this name. LIFF seemed to perfectly explain what we wanted to achieve. However, there was one problem with LIFF. Leeds International Film Festival was already using this acronym so we put an X for XML in front of LIFF and came up with XLIFF:

After about 10 months work we were at the stage where we had something which could be used for an interchange file format. The companies involved started using XLIFF in their work. We heard of interest from other people and started promoting the standard through conferences etc. It looked like everything was going well. It was then we got completely derailed by legal issues. At that stage we had not fully thought out Intellectual Property Rights and found that some of what we used within XLIFF was the intellectual property of a company who was not involved in developing the standard. This company wanted to support us but wanted us to acknowledge their ownership of part of the intellectual property. Lawyers from the big software companies got involved and we went into a wilderness for about ten months when we could not continue to work on XLIFF. We eventually found our way out of this by bringing the standard to OASIS and having all the companies involved agreeing that they were contributing their intellectual property.

OASIS was good for the development of XLIFF on many different levels. Engineers and developers treated work from OASIS very seriously while those involved in localization had always had difficulty persuading them of the importance of localization. OASIS saw our area as something that was important to them. They presented at localization conferences and worked with organizations from the localization industry. They also made connections between what their committee was doing in one area and what the XLIFF committee was doing. A good example of this was when the DITA and XLIFF committees joined forces to look at how localization should be done using DITA. One of the hallmarks of OASIS technical committee is that people working for competitors often find themselves on the same technical committee. I am not sure whether they ever had people Coke and Pepsi working on the same committee but the XLIFF team includes people who worked for companies that competed in the software industry, in localization and in language technology. Within the XLIFF team these people all worked together in a very positive and respectful way.


Lack of end user involvement in the design of technology is a pretty common fault. Language technology is often developed without any involvement, input or feedback from translators. This is the case with XLIFF and it is one area where I think we should try and change. We have people with a wide variety of experience and some of them are linguists. However, the experience being brought to the table is mainly from an engineering perspective and I think the standard would have benefited from professional translators helping to create it. So far, feedback from translators has been through articles they have published and presentations at conferences. All of this is good but it would be better if their involvement was more direct.

I did not join the group of companies meeting to look at XLIFF until the second meeting. Only Tony Jewtushenko of Product Innovator has been involved since the very start. For most of this time he has done an excellent job of chairing the committee. Recently he has been co-chairing with Bryan Schnabel of Tektronix. The committee has had some people come and go but it is made up of people who have made a lengthy and strong contribution to improving language technology. It is good to recognize this now that we are at the stage where XLIFF will soon become an OASIS standard. The good news is that these people seem intent on continuing.

Share:

Back to the Blog

Back to Top