April 2007 Archives

I wanted to continue my previous entry on the uses of machine translation. I spoke with Walter Hartmann, who has a long experience on MT. According to Walter, the idea of so-called “post-editable MT”, where the first translation is produced by machine and then edited by a human translator, faded quite much around 2001. This method is still used, as indicated by the continuous use of e.g. Personal Translator, PROMT, Systran, Tradissimo etc.

What seems to be increasing is the growing integration of TM and MT. In this approach, the texts to be translated are first run against TMs or translation memories for full or almost full matches, and then the rest is translated by MT, using a special dictionary which contains the special terms in the text. Even some of PC-based programs are integrating TM to a certain extent, and can be used with translation memories from TM.

This integrated approach definitely brings efficiencies and improves consistency through automation. Naturally, the better the original text, the better the translation – but that is actually true also for human translation, which does not eliminate the GIGO principle (garbage in, garbage out). Especially if you are palnning to have a text translated into multiple languages, it is a very good idea to pre-edit the text before sending it to translation.

It all brings us back to the importance of PROCESS, something Leonor has spoken so many times about, and which I absolutely agree on. Translation is not a separate task, it is part of the content management process. Managing multilingual content requires a well-designed process, otherwise you will end up in a costly chaos, from which even the best translators cannot save you!

PS. According to Walter, although it has been a while since the last update, the 12th edition of the Compendium of Translation Software http://www.eamt.org/compendium.html still has the most complete listing of MT programs available.

The integration of content and translation management workflows has a great deal of value for globalization projects. And as we've discussed, there are various market approaches to streamlining these increasingly complex processes. With the announcement of SDL International's intended acquisition of Tridion (set to close by end of May,) buyers officially have an additional approach -- translation and Web content management under one roof.

In this case, the opportunity is clearly for marketers who struggle to meet growing corporate and consumer demand for a multi-site, multi-lingual Web presence that drives revenue and protects brand (for the former) and delivers localized customer experiences (for the latter.) The time is right for this marriage, as globalization continues to climb toward the top of the CIO's "must-have" strategy list.

SDL and Tridion are undoubtedly headed toward a cohesive integration of their respective TMS and Web CMS technologies, which makes a great deal of sense for those organizations wishing to standardize on one platform for Web site translation and management. As we would expect, API-level workflow integration is at the top of the priority list, according to executives from both companies. There's quite a bit of potential for more, when one considers the ability of SDL's Author Assistant to enhance the value of content at its source, i.e. during content creation, as well as the power of Tridion's Communications Statistics module to drive process improvements based on data culled from user activities. Safe to say it will be interesting to watch the evolution of this combined product line for its impact on the Web content lifecycle.

As we've seen in the ECM and BPM suite market, the trend toward vendor consolidation changes the landscape dramatically and spurs the inevitable "suite versus best-of breed" debate. Within the globalization market, we expect this acquisition to follow suit -- after all, the marriage crosses the "dotted line" by solidifying the value of content and translation management integration.

At the end of the day however, the buyer defines the purchasing decision that makes the most sense, based on the most pressing -- or painful -- business requirements. As it stands now, Tridion will be a separate division within SDL and operate autonomously. R5 will be sold as a module within the SDL product set and renamed SDL Tridion R5. In parallel, SDL TMS will be sold as a Tridion module.

In effect, this strategy leaves decision-making in the hands of the buyer, as it should be. Hence, the immediate goal for this marriage is to demonstrate just how compelling the promise of a "total solution" will be. The CMPros community is already weighing in on the potential; Gilbane readers: join the conversation! We'd like to continue this discussion with your feedback.

Machine Translation (MT) was supposed to be the great solution for our translation needs. But how big a role does it have in the current translation industry? As far as I can see, in producing translations MT still has a very small role in the automated translation workflow, which relies much more on translation memories (TMs). But when we move to reading texts in foreign languages, such as web pages, MT has a much bigger role. The "Translate this" button provided by Google and many other search engines has a lot of users.

So at least for now MT is used primarily for getting a gist translation of a web site or a document. In producing a translation to be published, TMs and human translators still rule - perhaps with the exception of very strictly regulated or simplified texts which can be translated correctly enough by machine.

The reason is simple: the result from MT is still mostly so bad that a human translator can produce a full translation faster than what it would take to edit a machine-translated text.

Having said that, I do believe that major changes in MT are about to happen. I refer to speech recognition. In around year 2000, companies tried to build a Hal, which would understand not only what you said, but what you meant. Well, I have known my husband for almost 30 years, and I still do not always understand what he means, so that would have been a tall order indeed. Now, however, we have excellent applications based on recognizing a limited vocabulary - think of the times you have called an 800 number and talked to a machine which has understood you quite well.

MT will find more and more uses also in producing translations, when it will limit itself to a certain topic which it handles well, and combine efficiently with TMs. For gist translations, when one wants to understand e.g. the idea on a web page, MT is already quite useful.

This was one of my favorite quotes from Sun's April 10th presentation at Gilbane San Francisco, titled Managing Content Globally: What Works, What Doesn't. Given by Jed Michnowicz, Engineering Lead, and Youngmin Radochonski, Globalization Program Manager, the presentation opened the LISA Forum on Day 1 of the conference.

Jed and Youngmin nailed it when they defined three key components of a global content platform: content management, translation, and delivery. As they outlined the struggles of legacy challenges in all three areas, a pattern of checklist items for the audience quickly surfaced. Lack of metadata. "Siloed" mindsets, workflow, and content repositories. Static Web server content delivery. Inconsistent messaging. Slow time to market. Cost overruns. As moderator, it is always interesting to scan the faces in the crowd for reactions. During this part of the presentation the response was palpable: "got that, got that, and yes, definitely got that."

They also nailed it when they moved to the "here's the good news" part of the presentation. Global awareness throughout the organization. Process alignment and consistency. Separation of content from presentation. Translation memory management and sharing. Integrated content and translation workflows. Automated, Web services-based content distribution. They described what is most definitely a "Level 2+" integration from a technology perspective. At this point, the audience response was equally palpable: "want that, want that, and yes, definitely want that."

Wrapping up the success story with lessons learned (according to people, process, and technology categories; be still my heart!) Jed and Youngmin also noted that Sun, like most organizations, is still learning. Some of the questions they posed -- which we will continue to explore on this blog -- included:

What takes precedence when solving for people, process and technology?
What is the proper globalization strategy and who defines it?
Can a single solution work for everyone?

On behalf of The Gilbane Group and LISA, we thank these excellent presenters for a job well done. This presentation will be available here this week; check out one of my other favorite quotes emblazoned on the t-shirt on the last slide.