Translations - Article

Why Translation Control Matters

Most platforms support multiple languages, but many treat translation purely as a technical function. In practice, organizations often need much more control than simply converting text between languages.

Different industries frequently use different terminology to describe the same concept.

For example:

• Corporate learning environments may prefer Training or Programs
• Higher education institutions may prefer Curriculum or Education
• Industry-specific training providers may rely on terminology that is familiar within their profession or regulatory environment

Small differences in wording can significantly influence how intuitive and credible a platform feels to its audience.

The Translations configuration in Eurekos allows administrators to adapt the platform to their preferred terminology while maintaining consistency across the entire interface.

This capability supports:

• Multilingual learning platforms
• Industry-specific terminology
• Corporate branding alignment
• Customer training environments
• Regional language customization

In other words, the Translations feature allows the platform to communicate in the language and vocabulary your users expect.

What Translations Controls

Translations applies to platform-controlled text strings.

A string may be:

• A single word
• A short phrase
• A headline
• A short interface message
• A system notice

Examples include:

• Menu items
• Interface labels
• Standard system descriptions
• Notifications shown within the platform
• Selected site-wide text elements

These strings form the language that shapes how users experience the platform.

What Is Not Managed Here

The Translations area does not control user-generated content.

Examples of content that are translated elsewhere include:

• Course descriptions written by administrators
• Questionnaires and assessment content
• Tags and vocabularies created in taxonomy features
• Learning materials created within activities

These are managed within their respective features.

The Translations configuration focuses specifically on system interface language and standard platform text.

Translation Tabs

The Translations area is organized into several tabs managed from Settings → Translations.

Not every tab will be relevant for every environment. Most administrators primarily work within the Interface and Site tabs. The Mobile App and Translation Models tabs are generally relevant only when those features are enabled as part of the platform configuration.

Interface Tab

The Interface tab is where administrators manage the majority of translation strings used across the platform. This tab displays a searchable list of all available interface strings.

Each row represents a system text element that can be translated or customized.

Administrators can:

• Search for strings
• Open individual strings for editing
• Modify text across multiple languages
• Export strings for external translation
• Import translated strings back into the platform

This tab therefore acts as the central terminology control layer for the platform interface.

How The String List Works

The string displayed in the list reflects the administrator’s current language preference.

For example:

• An administrator using English will see the English translation of the string
• An administrator using German will see the German translation of the same string

This makes it easier to locate and review translations in the language the administrator is working in.

However, the edit interface allows administrators to modify all active languages simultaneously.

Editing A Translation String

Selecting the More options (…) menu for a string opens the translation editor.

The editor contains:

• The Original text string
• Context information describing where the string is used
• Input fields for each active language

This allows administrators to supervise and edit translations across languages in a single view. Because a single string may appear in several parts of the interface, changes should always be reviewed after editing.

Searching For Strings

The search field allows administrators to locate specific strings by entering part of a phrase. For example, searching for a term such as “Time” may return several results. This occurs because the same wording may appear in different contexts across the platform.

When editing search results, administrators should:

• Confirm that the string represents the correct context
• Review where the string appears
• Verify the change across relevant pages after editing

The translation system is powerful but requires careful review to ensure that edits affect the intended interface elements.

Dynamic Tokens And Variables

Some strings contain dynamic tokens, which represent variables automatically replaced by the platform.

Examples include strings such as:

• %time ago
• Created by !author time

These tokens are placeholders that the system fills with dynamic data. These include time values, author names, dates, activity titles and many more. 

When editing strings containing tokens:

• The token syntax must remain intact
• The token may appear in different word order depending on language
• The surrounding sentence should still read correctly once the value is inserted

If a token is changed incorrectly, the system will still function, but the message may appear incorrectly to users. The Original text field displayed in the editor allows administrators to restore the original token structure if necessary.

Import And Export Of Translation Strings

Large platforms often manage translations through structured workflows rather than manual editing of individual strings. For this reason, the Interface tab supports import and export functionality.

This allows administrators to:

• Export strings for translation
• Collaborate with translation agencies or language specialists
• Review terminology with internal teams
• Re-import translations in bulk

This workflow is particularly useful for multilingual deployments.

Translation File Formats

The export feature typically supports standard formats used in software localization.

Two common formats are:

• PO Files (.po)

PO files are a widely used translation format originating from the GNU gettext system. They are commonly used in software localization workflows and supported by many professional translation tools.

PO files maintain structured relationships between:

• Original text
• Translated text
• Context information
• Dynamic tokens

Because of this structure, they are ideal for professional translation agencies using localization platforms or translation memory systems.

• Excel Files (.xlsx)

Spreadsheet exports provide a more accessible format for internal teams who may not use specialized translation tools.

Excel workflows allow administrators or translators to:

• Review translation strings in bulk
• Adjust terminology across many entries
• Collaborate with internal reviewers

This format is often used when translation work is handled internally rather than through a localization vendor.interface design differ.

You’re absolutely right — that is critical operational information and should definitely be included. Import behavior can override existing translations, and plural forms are particularly sensitive because they control how the system renders quantities in different languages. If administrators misunderstand this, they could unintentionally overwrite translations or break plural handling.

Import Behavior And Plural Form Handling

When importing translation files, it is important to understand how the platform handles existing translations and plural forms. During import, administrators can choose how the system should treat the uploaded strings.

Depending on the selected import option, the system may:

• Overwrite existing translations with the imported values
• Add new strings while keeping existing translations unchanged
• Preserve plural form structures used by the platform

Because of this behavior, administrators should carefully review the import configuration before confirming the upload.

Overwriting Existing Strings

If the import mode allows overwriting, the system will replace the existing translations for the affected strings with the values contained in the uploaded file.

This means:

• Existing translations may be replaced entirely
• Previous wording changes may be lost
• Translations edited directly in the interface may be overwritten

For this reason, it is recommended to export a backup of the current translations before performing large imports.

Plural Form Handling

Some interface strings support plural formats, which allow the system to display different wording depending on numeric values.

For example:

• “1 Course”
• “5 Courses”

Plural handling differs between languages and is managed through specific plural syntax in translation files.

When importing translation files:

• The system preserves existing plural structures
• Imported translations must maintain the correct plural format
• If plural formatting is removed or altered incorrectly, the message may render incorrectly

The interface provides an information message indicating when existing plural forms are preserved during import.

Administrators should therefore:

• Avoid modifying plural syntax unless they understand the format
• Ensure translation providers keep plural structures intact
• Test pluralized strings after large translation imports

Recommended Workflow

To avoid unintended overwrites or formatting issues:

1. Export Current Strings Before Importing Updates
2. Review The Import Mode Carefully
3. Ensure Plural Forms Are Preserved In The Translation File
4. Validate Important Interface Areas After Import

Following this process ensures that translation updates remain controlled and predictable.

Yes — it fits very naturally if placed immediately after the Import / Export section. That is the point where administrators start thinking about managing translations at scale, and governance becomes relevant.

Translation Governance For Multilingual Platforms

When a platform operates in multiple languages, translation management becomes an ongoing operational responsibility rather than a one-time configuration task.

Administrators should therefore treat translation management as part of platform governance, ensuring that terminology, tone, and structure remain consistent across languages and over time.

In practice, this involves coordinating several aspects of the platform:

• Interface translations
• System emails
• Site-level messaging
• Mobile app terminology (if used)
• Industry-specific vocabulary and branding

Without a coordinated approach, multilingual platforms can gradually develop inconsistencies where similar concepts are described differently across features, languages, or communication channels.

Establishing Terminology Consistency

Organizations often maintain a preferred terminology set to ensure that key concepts are described consistently across the platform.

For example:

• A corporate academy may consistently use the term Training Program instead of Course
• A certification provider may standardize on Certification Pathway
• A customer training platform may align terminology with product documentation

Using the Translations interface to enforce these choices ensures that terminology remains consistent across navigation menus, system notices, and standard interface text.

Coordinating Translation Workflows

When multiple languages are supported, translation updates are often coordinated through structured workflows.

A typical governance process may include:

1. Export Translation Strings For The Target Language
2. Review Or Translate The Strings With Internal Language Owners Or External Translation Providers
3. Validate Terminology Against Brand And Industry Standards
4. Import The Updated Translation File Into The Platform
5. Review The Platform Interface To Confirm The Translations Render Correctly

This structured workflow ensures that updates remain controlled and that translation changes do not introduce unintended inconsistencies.

Maintaining Quality Over Time

Translations should also be reviewed periodically as the platform evolves. New features, interface updates, and terminology changes may introduce new strings that require translation or review.

Regular review helps ensure that:

• Terminology remains consistent across features
• Newly introduced interface elements are translated appropriately
• Changes to brand language or organizational vocabulary are reflected in the platform

Organizations operating global training platforms often schedule periodic translation reviews alongside other platform governance activities.

Mobile App Tab

The Mobile App tab contains translation strings used in the mobile application. While the mobile app has a different interface design, it still relies on many of the same core concepts as the web platform.

• When organizations customize terminology in the main platform, it is good practice to review the mobile app translations to ensure consistency
• Maintaining consistent wording across web and mobile environments helps users recognize features and navigate the platform more easily

If the mobile application is not enabled in the environment, this tab may not be relevant.

Site Tab

The Site tab manages selected site-level text elements. These are typically texts that appear in broader areas of the platform rather than within specific interface components.

Examples include:

• Maintenance mode messages
• Cookie notices
• Footer text
• Copyright statements
• Branding slogans

Because these texts appear across multiple pages, they have a strong influence on the overall presentation and branding of the platform. Like the Email Sending configuration, the Site tab allows administrators to select a language and edit the content separately for each language.

Adding A New Language

New languages can be activated through Settings → Regional and Language → Languages.

Once activated, the language becomes available throughout the platform. However, enabling a new language can introduce significant governance considerations.

Administrators may need to review:

• Interface translations
• Site text
• System emails
• Mobile app strings
• Terminology consistency

For this reason, new languages should generally be enabled when there is a clear plan for maintaining translation quality.

Translation Models

The Translation Models tab supports advanced translation workflows and it is an opt-in option. This functionality is used for translating structured content such as course content or H5P components through machine translation models.

Translation models are connected to Google Cloud storage and translation datasets used for training translation models. Because training translation models involves external services and costs, this feature is generally used in specialized implementations where large volumes of content require automated translation.

Organizations interested in using translation models should coordinate configuration with the Eurekos Service Desk.

Yes — this section can definitely be improved. The current version reads a bit like internal notes, while the article should present it as clear operational guidance that explains why the workflow exists and how it works. We can also structure it so administrators understand that this is an advanced capability used for training machine translation models, not something used in everyday translation tasks.

Import Training Data

Both Translation Models and Translation Datasets support an Import Training Data workflow.

This capability allows administrators to generate structured translation data that can be used to train machine translation models for learning content. Instead of translating content manually each time, organizations can train models using previously translated material so that future translations become more accurate and consistent.

The process works by comparing two versions of the same course:

• A source course in the original language
• A target course containing the translated version

The system analyzes the relationship between these courses and extracts aligned translation segments that can later be used as training data.

How The Training Data Import Works

When importing training data, the system performs the following steps:

• The Source And Target Courses Are Compared—The platform analyzes the content structure of both courses and extracts translation pairs from the corresponding elements
• Course Structure Must Match—For the dataset to be created correctly, the source and target courses must contain identical structure. This ensures that text elements can be matched accurately between languages
• Only Courses With A Defined Language Are Available—The course selectors only display courses that have a language configured in Settings → Regional And Language
• Large Courses May Be Processed Asynchronous—If the source course contains a large amount of content, the import process may be queued and processed asynchronously
• Completion Notification Is Sent By Email—Once the dataset generation process is completed, the initiating user receives an email notification
• Generated Translation Data Is Uploaded To Google Storage—The resulting translation dataset is stored as a TMX file and uploaded to the configured Google Cloud storage bucket used by the translation model configuration

TMX Translation Segments

The generated dataset is structured using the TMX (Translation Memory Exchange) format, a widely used standard in translation and localization workflows.

Within the generated TMX file:

• Each text widget is treated as a separate translation segment
• Each button widget is also represented as its own translation segment

These segments form the training data used for translation models.

The TMX file can also be downloaded and reviewed, allowing administrators or language specialists to validate the extracted translation pairs before using them in model training workflows.

Using The Dataset For Model Training

Once the dataset has been generated, it can be imported into Google Translation Datasets, where it is used to train or refine machine translation models. These trained models can then be used to support automated translation of learning content across languages.

Because this workflow is designed for model training and large-scale content translation, it is typically used in environments where organizations manage large multilingual content libraries.

For everyday translation tasks such as editing interface wording or standard platform text, administrators should use the Translations configuration instead.

Best Practice Recommendations For Managing Translations