Prefill and fallback languages - Learn the ins and outs of these useful features

· 8 min read
Post featured image

Edit on 22nd July 2021: Autofill is now called prefill. Check out this changelog entry here.

Localization and handling a vast amount of translations can be a challenge for even the most experienced professionals. This is even more true, when you are dealing with content that will be localized for pluricentric languages.

Pluricentric languages are languages where one common standard and several regional variants coexist. One example is English with British, American, Australian, and many more variants. A second example would be the one of my mother tongue, German, with variants in e.g., Austria, Swiss, and Northern Italy.

So how can you localize your content for so many different languages and variants easily? Today I will introduce you to LingoHub's autofill prefill and fallback language features and explain what they can do and when you should utilize them.

Here's a quick overview of the topics I will cover, feel free to jump right into a section, in case you are familiar with some content already. Or keep scrolling to view them all.

What is Prefill?

Prefill is one of LingoHub's powerful translation tools. It is a feature which fills empty text segments with content of matching texts from your translation memory, machine translations, or from existing languages.

You can choose to prefill empty text segments with two sorts of translation memory matches:

  • Fuzzy matches: These are text matches with a percentage of more than 80% or 90%, depending on your settings.
  • Exact matches: These are text matches with a percentage of 100%.

The prefill feature will search for matches in your translation memory. Depending on your settings, a translation memory can be organization- or project-wide.

You can also select to prefill your empty text segments with machine translations, or an existing language.

For a more detailed guide on the prefill feature, visit our Help Center article.

What is the fallback language feature?

A fallback language is a language that will be used when a text segment is empty upon export, e.g., due to a missing translation. When activated, the fallback language feature ensures that empty text segments will be filled with alternative content from another language.

With LingoHub you can easily define fallback languages for every target language. Should your fallback language also be empty, the fallback language of your fallback language will be used. LingoHub automatically searches for fallback languages until the source language has been reached.

This may sound more complicated than it is, so let's look at an example. German (de-DE) and Austrian German (de-AT) use identical words, apart from a few exceptions. So when we translate coffee into these languages, this is the result:

Text segment Language Fallback language
Kaffee de-AT (= target language) -
Kaffee de-DE (= target language) -
coffee en-US (= source language) -

As you can see, the translation for coffee is the same in both target languages. This is perfect for the use of the fallback languages feature.

To activate the fallback language feature, we set the fallback languages for each target language and remove the Austrian German translation:

Text segment Language Fallback language
de-AT (= target language) de-DE
Kaffee de-DE (= target language) en-US
coffee en-US (= source language) -

Upon exporting, LingoHub will automatically detect that the text segment for de-AT is empty. It will check for the fallback language in German, see that it is not empty and export the following:

Text segment Language
Kaffee de-AT
Kaffee de-DE
coffee en-US

In a section below, I will illustrate how you should proceed when there are different translations for language variants. For a detailed guide on the fallback language feature, visit our Help Center article.


We have established the basics of the prefill and fallback language features. Let's talk about when to use them.


When to use Prefill

Prefill is excellent as a starting point for your translators. Whenever you want to translate your content to a new language, simply use the feature to prefill the empty segments. You will never have to start with a blank canvas again.

Your segments can be filled with:

  • fuzzy matches from your translation memory
  • exact matches from your translation memory
  • machine translations from DeepL, Amazon Translate, and Google Translate

Please be aware that this cannot replace professional translations, and quality assurance has to be performed.

However, it is still a great way to check your localization. As you may know, translations can vary in length. The prefill feature can also be used to check the length of your text in different languages and help you with localizing your user interface for different languages.

When to use the fallback language feature

Fallback languages can boost your translation efforts drastically. Especially with pluricentric languages or customer specific content, the fallback language feature can shine.

As I have shown you before, for pluricentric languages one base language (e.g., en-GB or de-DE) exists, which contains texts for all text segments.

Text segments of the language variants (e.g., en-AU or de-AT) can stay empty if:

  • the base language is used as a fallback language
  • both share the same translation

You only have to change text segments that are different in the language variants.

Let's look at an example in the English language.

In the first text segment example, we want to localize a greeting in American, British, and Australian English.

Text segment Language Fallback language
G'day en-AU (= target language) en-GB
en-US (= target language) en-GB
Hello en-GB (= source language) -

As you can see, the greeting for British English and American English is identical. Therefore, the text segment for en_US can be left blank, LingoHub will use the fallback en-GB. Australian English features a different greeting compared to the others, therefore it must not be left blank but instead filled with the regional variant "G'day".

Let's look at another example.

For the word color, two different spellings exist. Our table will look like this:

Text segment Language Fallback language
en-AU (= target language) en-GB
color en-US (= target language) en-GB
colour en-GB (= source language) -

In this case, the fallback language is only correct for Australian English. For American English, we must enter the text "color".

In the last example, no differences between the three languages are present. As described earlier, this is the most common case to use the fallback language feature. You can leave two text segments empty and use the fallback option.

Text segment Language Fallback language
en-AU (= target language) en-GB
en-US (= target language) en-GB
Apple en-GB (= source language) -

The examples above are very basic, but regardless of your text segment complexity, you will still benefit from implementing the fallback language workflow. Especially, when you localize to pluricentric languages.

The same workflow can be applied when dealing with company specific content. If your content is identical except for specific brand names or corporate wording, the fallback language workflow can also be applied.

Note: For uploads, use the fallback language text segments with caution.

We advise you to not upload exported text segments again, after having used the fallback language feature. When uploading, your empty text segments will be overwritten. There is no setting to change this behavior.

The example below illustrates what would happen, when you upload your target languages back to LingoHub.

  1. Data before the export:
    Text segment Language Fallback language
    de-AT de-DE
    Apfel de-DE en-GB
    Apple en-GB -
  2. Export data/import data
    Text segment Language
    Apfel de-AT
    Apfel de-DE
    Apple en-GB
  3. Data after import
    Text segment Language Fallback language
    Apfel de-AT de-DE
    Apfel de-DE en-GB
    Apple en-GB -

As you can see, LingoHub will automatically overwrite your previously empty text segment in the de-AT language. The fallback language workflow will not work anymore for the de-AT variant as it has been filled with content.

We therefore advise you to only export data when using the fallback language feature, unless you are sure you want your import to overwrite previously empty text segments with fallback language translations.

Conclusion

With this blog post, I introduced you to the prefill and fallback language features and illustrated the different use cases for both. When used right, they can help you push your translation to the next level. Be sure to read up on the mentioned Help Center articles as well, as they feature more detailed information and step-by-step guides on how to set up the features properly.

Also stay tuned and check back soon! In a future blog post, I will tell you more about all translation tools available with LingoHub and how to best utilize them.


Are you not using LingoHub yet? Try it out; it's entirely free for 14 days and all features are available to you during your trial. Feel also free to book a demo, or reach out to us anytime, simply contact our support.

Clock calendar notifications in harmony

Start your 14-day free trial