Yes, I want my consultation

Translation Productivity Tool You Should Try Gets Even Better

As has been the case with all recent major updates, OmegaT 2.5.4 released last week introduced a few very important improvements. Let’s take a look at some of them to find out how they can make a difference in your daily work as a translation professional.

Automatic Number Replacement

Perhaps, the greatest improvement is having the numbers replaced automatically when you insert a fuzzy match with numbers that differ from those in the current source. This feature is available in other translation memory tools, and you’ve been probably longing for this feature in your favorite translation program just as I have. Here you go! You can find this new option under Options > Editing Behaviour > Attempt to convert numbers when inserting a fuzzy match.

I remember a German to Russian translation project from the last year where not being able to replace the numbers was particularly painful because the source file included so many segments that were different only in numbers. I decided to return to this translation now to test-drive the long-awaited feature.

I started by checking the new feature on standalone numbers:

1 Niederdrucküberwachung 4 Kreis 1

2 Niederdrucküberwachung 5 Kreis 2

3 Niederdrucküberwachung 6 Kreis 3

After translating the first phrase, I proceeded to insert the translation into the other two. OmegaT replaced the numbers correctly in the third segment, but in the second one, it got the order wrong. Based on this test and other translations I checked this feature on, I conclude that this feature works very well, but it might not be perfect. “Attempt” in the name is obviously there for a reason. So, don’t rely solely on it and keep checking your translations for number errors with your favorite translation QA program such as CheckMate.

Another situation is where numbers are glued to words and it might be a difficult one. For example, the popular program SDL Trados wasn’t able to replace this type of numbers back in the old days:

Strom L1

Strom L2

Strom L3

I found that at this point, OmegaT doesn’t replace this type of numbers.

To summarize, with this new feature available, I estimate that I could save at least 20% of time on that German project—quite an astonishing result for just one improvement. I know you are also longing to know whether this feature also applies to numbers to tags. Well, it doesn’t, at least for now. Here is an example:

Select <t0/>Yes<t1/> to restart the system.

Select <t2/>Yes<t3/> to restart the system.

In this case, when you insert a fuzzy match in the second segment, OmegaT won’t replace the numbers. This would be a very nice feature to have in a future version, just as the ability to replace entire tags, e.g. <x0> and <x1> by <t3> and <t4>.

I also hoped that this new feature also extends to the statistics, but it doesn’t. The Statistics function recognizes similar segments with numbers as completely new. And the Match Statistics function continues to recognize them just as fuzzy matches—not 100% matches as you’d expect.

Auto-Propagation in TTX Files

Another long-awaited feature implemented in OmegaT 2.5.4 is the ability to use auto-propagation with bilingual files such as TTX or TXML. In fact, we even suggested a workaround to avoid the underlying problem in a recent post about preparing TTX files. Previously, when you loaded a segmented file, OmegaT recognized all segments as alternative translations, which prevented auto-propagation of repetitions. You had to change each repetition manually. This is no longer an issue. If you load a translated or segmented TTX now, you can go ahead and change any repetition—all other repetitions will change automatically.

It’s my opinion that the current approach still has a few minor limitations. All segmented texts in your TTX file, including those where source is equal to target, appear as translated in the Editor pane, which might be a little bit confusing for those of you who use OmegaT on a regular basis. If you delete any of these translations, it will reappear after reloading the project. These translations also display in the Fuzzy Matches pane, taking up valuable space. It would be great to see improvements in this area in future releases.

Auto-Propagation Now Also Works in TXML Files Processed with Okapi’s Filter

I also tried this new feature with TXML files. They must be segmented as well, and Okapi’s Wordfast Pro TXML filter used to have the same problem as above. OmegaT, however, provides a native Wordfast TXML filter, which didn’t have this problem. So, if you are using Okapi’s filter, you are also going to benefit from this new improvement. One reason to use Okapi’s filter is that because unlike the native one, it doesn’t hide the tags in the beginning of the segments, which makes it impossible to move the tags in the translation as may be required. Here is an example:

Wordfast: {ut1}XYZ{ut2} support staff are well-mannered.

Native filter: XYZ<u0/> support staff are well-mannered.

Okapi’s filter: <x1/>XYZ<x2/> support staff are well-mannered.

In this segment, the direct word order of the Russian language requires that we put the translation of “support staff” in front of “XYZ,” hence having the ability to move the opening tag is crucial. Okapi’s filter provides this ability by displaying both the opening and closing tag while OmegaT’s native filter doesn’t.

Changing Save Interval

You can now set the automatic save interval by selecting Options > Save. It used to be 10 minutes, but now you can select a shorter interval if you want. For example, a shorter interval can help you retain more translations in case OmegaT or your PC crashes. Additionally, when several translators are collaborating on the same project and you are frequently copying project_save.tmx files to a network folder for sharing the translations, a shorter interval can increase the speed of making those translations available to the team.

With all the great improvements happening with OmegaT recently, it’s really wise to give it a try if you still haven’t. Go to OmegaT’s download page for the latest version. Or check out the marvelous OmegaT community site on Yahoo, which is a source of incredibly fast and comprehensive support and already provides a wealth of information about using the program.


  • Cassy says:

    Very informative post! Actually, I haven’t tried this before. With all these improvements that you’ve mentioned, I think it’s worth to try.

  • Marco Cevoli says:

    I’m very happy that someone else realized how important is to have this feature in OmegaT. We thought it was so important that we even financed this development. In fact, with this comment I simply wanted to let your readers know that OmegaT lets you contribute to its development: if you think some feature is missing, just let the developers know. If they deem this feature important too, you can go on by financing (totally or partially) the corresponding programming. This is the power of collaborative software and of open source. 🙂
    Congratulations for your blog, it’s very interesting. Keep on the good work.

  • Hi Marco,
    We thought it was so important that we even financed this development
    You did?! Wow, that sure is a great example for the rest of OmegaT users to follow. Honestly, I’ve been thinking about donating to the team for a long time, but simply procrastinated. I will now start doing that on a regular basis.
    And it’s definitely a great feature. As an ex-user of Trados, I’ve been missing it. The project I’m talking about in this post was really a pain until I figured out that I could use Trados to extract the unknown segments, translate them in OmegaT, and then use the TMX to translate the final files in Trados.
    Your advice about micro-financing development of OmegaT’s features is excellent. It’s so easy to talk yourself into not donating when you think about fairly large amounts of money—just like I did. To avoid that, we should simply donate whatever we can on a regular basis, and in the long run, it’s going to make a difference even if indivudual amounts were extremely small.
    Thank you very much for your comment and tweets!
    With best wishes,

Add comment

About the Author

Roman Mironov
Roman Mironov
CEO & Founder

As the founder of Velior, Roman has had the privilege of being able to turn his passion for languages into a business. He has over 15 years of experience in the translation industry. Roman has helped dozens of clients increase sales by making their products appealing for speakers of other languages.