How can you translate dominKnow ONE e-learning courses into several languages without losing control of translation approvals, project relationships, variables or learning outcomes? The answer begins with the difference between a multilingual project and separate translated projects. In dominKnow ONE, that choice shapes how language editions are governed, reviewed, packaged and maintained. The words matter, but the relationships between those words matter just as much.
This dominKnow ONE localization guide explains administrator language settings, AI-assisted course translation, manual export and import, XLIFF 1.2 and 2.1, Excel and XML workflows, language-specific terminology, system variables, multilingual review and publishing quality assurance. It focuses on practical decisions that instructional designers, translators, course owners and training administrators must make when they deliver professional learning across languages.
Official dominKnow guidance documents enabled and default language settings, per-language translation file formats, one-project multilingual publishing and separate-project translation strategies. Its language-management materials also describe AI assistance and organization-wide governance. These are useful capabilities, but they are not substitutes for checking technical identifiers, assessment decisions, accessibility, local policy authority and the target learner’s ability to complete the course.
The useful answer first
Treat every dominKnow localization as a controlled learning revision. Define the master, select approved locales, determine whether the audience needs a unified multilingual package or regionally separate courses, and inventory the content beyond visible paragraphs. Translate the intended actions, interactions, feedback and assistance routes; then review the result inside the working course and under the same publication conditions the learner will encounter.
The best workflow does not ask the translator to guess a compliance rule or the platform administrator to adjudicate language nuance. It gives every important decision an owner. A source specialist determines what must remain true; a language specialist makes it understandable; an instructional designer validates learning; and a release owner verifies that the correct edition reaches the intended audience.
Understand the dominKnow ONE model before translating
dominKnow ONE combines authoring, translation and project management in a way that can make course maintenance more coherent. An organization may keep multiple languages associated with one project or use separate language-specific projects. The first pattern suits material that must remain recognizably the same course. The second can suit local variations with distinct ownership or substantially different release requirements. The platform’s support material explicitly discusses both approaches; neither should be selected by habit.
Administrator choices influence what authors can do. The organization can control which languages are available and which are selected by default; settings can also specify file formats used for different target languages. This prevents course authors from inventing inconsistent language selections and file handoffs, but it makes the configuration an important part of course governance. A language list is a decision about what the organization can responsibly support, not a claim that every possible locale is already quality-assured.
The export routes include formats such as Excel, XML and XLIFF, with documented XLIFF versions including 1.2 and 2.1. The right route depends on translator tooling, markup preservation, the shape of the project and the review process. Format compatibility should be piloted with a representative course before a language agency receives the full workload. A file that opens in a CAT tool is not necessarily a file that will reimport correctly.
AI-assisted translation can reduce the first-draft burden. dominKnow describes an AI Translator integrated into its project workflow and the ability to route output through existing review and approval processes. A responsible team uses this capability where it saves work, then assigns expert checks according to risk. A short internal orientation and a safety-critical certification do not warrant identical acceptance standards.
The exported text can contain far more than page body copy. dominKnow’s file guidance describes course, module, learning-object, page and element context, alongside text from actions, variables, glossaries and links. Some automatically supplied player strings are available in many languages as system variables. This is a reminder to map content by function: a prompt, a value, an identifier and a feedback message do different jobs even if they look like strings.
Consider a target-language course from discovery to completion. A learner must identify the correct version, navigate its sections, understand the questions, receive unambiguous feedback and obtain the right completion record. If these steps do not work together, passing a translation-file check is insufficient. Quality belongs to the whole learning experience, not only to the exchange document.
Worked example: translating customer-support escalation training
Imagine a global customer-support course with three branches: an ordinary request, a security concern and an urgent service outage. Each branch asks the learner to choose a response, read customer-facing language, escalate to an approved channel and understand what information should be recorded. The English master is already reviewed, but the translated editions must work for teams with different communication norms and region-specific escalation arrangements.
First map the invariant decisions: a security concern must never be treated as a routine service ticket, a customer must not be promised an unsupported fix, and private information must not be copied into an unauthorized channel. Then separate the local facts: contact systems, operating hours, jurisdiction-specific disclosure language and escalation teams. A translator can make a polite phrase idiomatic, while the approved business owner must establish which hotline or procedure is valid.
In dominKnow, select one multilingual project if the branches and learning objectives remain consistent across regions. If business units require different procedures and separate review authorities, decide whether language-specific projects offer a more reliable boundary. Configure languages at organizational level, choose the suitable translation route and export a short pilot first. For an Excel handoff, examine the Current Text, Translated Text and contextual columns rather than asking translators to fill cells blindly.
Translate the visible dialogue and the invisible interaction strings together. A pop-up warning, a branch condition, a variable displayed inside an answer and a feedback message may all affect what the learner understands. Protect markup and identifiers when translating structured export files. Have reviewers complete each branch without seeing the English master, then compare the results against the approved competence model.
The release decision should be supported by three kinds of evidence: linguistic explanations for important terms, functional tests showing that branches and scoring behave correctly, and deployment tests proving that learners receive the intended language and completion data. Keep a short record of open issues and resolved terminology disputes, because a course is not static once it has been released.
Twenty-eight practical dominKnow ONE localization decisions
1. Enable only governed languages
Situation. An author sees a large list of potential translation targets and selects dialects for which no reviewers or help resources exist. This creates a promise the team cannot keep.
Method. Use the organization’s Languages settings to make support decisions explicit. Approve locales according to real learner demand, reviewer availability, appropriate terminology and the ability to maintain future updates.
Verification. Check the language chooser against the project’s published support policy. A locale should not become available merely because an automated service can produce plausible text. Here is the deeper principle: language quality cannot be judged from isolated strings. The author must understand what the learner is trying to accomplish, which course mechanism presents the information and what evidence would establish that the learner understood it. If those three elements are disconnected, a fluent translation may still create a wrong or unfair learning experience.
2. Configure default target languages
Situation. Every author repeatedly selects a different set of languages, and a required regional edition is sometimes forgotten.
Method. Set defaults centrally for the languages expected on ordinary projects, then permit justified exceptions. Treat defaults as a starting point, not as evidence that a target is already reviewed.
Verification. Open a sample project and compare its offered targets with the approved rollout matrix. Confirm that specialized teams can request legitimate additional language support. Think about responsibility before efficiency. A translation service can propose wording quickly, yet it cannot know which workplace rule the source owner is empowered to change. Bring factual disputes back to that owner and ask a language specialist to explain the effect of the phrase. That keeps convenient automation from becoming accidental policy writing.
3. Choose one multilingual project
Situation. A course teaches a shared universal skill, yet the team proposes one fully independent package for each language and then duplicates every revision.
Method. For content with shared assessments and equivalent structure, evaluate one multilingual project and its learner language selection. Keep localized assets distinguishable without changing the core lesson unintentionally.
Verification. Test how language switching works and whether the course reporting remains understandable in the destination LMS. The decision is about maintenance as well as initial translation. An effective review has a defined acceptance question. Instead of asking only ‘Does this sound good?’, ask whether the learner can perform the specified decision without seeing the original course. A short scenario, followed by an explanation in the target language, often reveals more than several rounds of cosmetic correction.
4. Choose independent language projects
Situation. A single master attempts to contain region-specific regulations, separate approval chains and differently sequenced training objectives.
Method. Use separate language projects when local course authority and independent release dates are essential. Record which core competences remain shared across the projects and which are genuinely local.
Verification. Have each owner identify the edition they are authorized to publish. Compare curriculum objectives so independent delivery does not quietly produce incomparable training claims. Protect the future maintainer as well as today’s learner. A course may look excellent now and still be difficult to update if the project contains undocumented local exceptions, overwritten wording and ambiguous file names. Preserve the small decisions that are likely to matter when the original author is unavailable.
5. Choose Excel rather than XLIFF when appropriate
Situation. A language reviewer is a subject specialist who can inspect a structured spreadsheet but has never used translation software.
Method. dominKnow’s Excel route exposes contextual columns and designated translation cells. Explain exactly where translated text belongs, how markup must be handled and which cells are structural.
Verification. Test one small import containing a formatted heading, interactive label and glossary phrase. Verify that a successful spreadsheet import leaves the page behavior intact. Separate the format test from the education test. A file can be valid XML or XLIFF while a sentence misstates the rule. A rendered page can look polished while a branch chooses the wrong outcome. Use distinct tests for import integrity, meaning, interaction, accessibility and delivery rather than letting one green check mark stand for all five.
6. Choose XLIFF 1.2 or 2.1 deliberately
Situation. A translation agency asks for XLIFF 2.1 but the project team assumes all XLIFF packages will behave identically in its existing tools.
Method. Check the version explicitly supported by each translator’s environment and the dominKnow target settings. Pilot inline tags, line breaks and variable placeholders through a complete round trip.
Verification. The acceptance test is the rendered course with correct interactions, not merely a file with an XLIFF extension. Keep a copy of the approved export revision. A learner encounters a sequence, not a list of phrases. They read a heading, interpret an instruction, make a choice and receive feedback. If one translated link or conditional message breaks that sequence, the remaining pages cannot compensate. Follow the task from discovery to completion and inspect the transitions.
7. Preserve XML CDATA and structure
Situation. A translator opens an XML exchange file and inserts translated words outside the expected CDATA region, breaking reimport.
Method. Provide format instructions and a structural validation step. Translate content only in the intended fields, keeping machine-readable boundaries, attributes and identifiers exactly where required.
Verification. Reimport a small sample and compare the target page to the source. Include an example with punctuation, quotes and a mixed-script identifier to expose encoding problems. Do not assume that content uniformity is always fair. People may need different examples or locally approved procedures to learn the same underlying competence. Preserve the shared objective while documenting legitimate variation; otherwise one supposedly universal edition may become incomprehensible or misleading to part of the audience.
8. Handle embedded text markup
Situation. The exported Current Text cell contains formatting markers around a warning, and a translator removes those markers as if they were disposable clutter.
Method. Explain which characters create styling or link structure, and which are the words to translate. Provide protected-tag handling in the translator workflow where possible.
Verification. Inspect the target warning visually and with keyboard navigation. The important test is whether the emphasized phrase still conveys the intended severity and supports the correct action. Here is the deeper principle: language quality cannot be judged from isolated strings. The author must understand what the learner is trying to accomplish, which course mechanism presents the information and what evidence would establish that the learner understood it. If those three elements are disconnected, a fluent translation may still create a wrong or unfair learning experience.
9. Translate pop-up actions
Situation. A course opens a modal after an incorrect decision. Its body text is localized but its close label and link instruction remain in the source language.
Method. Inventory action-related text, including pop-ups and dynamic links that may appear in separate export sheets. Translate the whole interaction, not only the visible default screen.
Verification. Trigger every decision path during review. Check whether a novice user can close the pop-up, read the reason for the mistake and continue without guessing. Think about responsibility before efficiency. A translation service can propose wording quickly, yet it cannot know which workplace rule the source owner is empowered to change. Bring factual disputes back to that owner and ask a language specialist to explain the effect of the phrase. That keeps convenient automation from becoming accidental policy writing.
10. Check variable-driven text
Situation. A feedback message inserts a variable that contains a team name or a number. The translator treats the variable expression as ordinary prose.
Method. Keep variable syntax intact and translate the surrounding sentence so it remains grammatical when the actual value appears. Test with more than one possible value.
Verification. Replace the sample variables with long names, short names and unusual grammatical forms. Confirm that the resulting sentence remains truthful and understandable. An effective review has a defined acceptance question. Instead of asking only ‘Does this sound good?’, ask whether the learner can perform the specified decision without seeing the original course. A short scenario, followed by an explanation in the target language, often reveals more than several rounds of cosmetic correction.
11. Review imported glossary terms
Situation. A glossary term reused from an older project arrives pretranslated, but the course gives the term a more specialized meaning.
Method. Do not accept inherited translations simply because the export contains text. Check definitions, audience and subject context against the current course.
Verification. Ask the reviewer to explain the concept without reading the bilingual entry. A correct mapping is one that supports the learner’s meaning, not just visual consistency. Protect the future maintainer as well as today’s learner. A course may look excellent now and still be difficult to update if the project contains undocumented local exceptions, overwritten wording and ambiguous file names. Preserve the small decisions that are likely to matter when the original author is unavailable.
12. Use built-in system labels carefully
Situation. A platform-provided Submit button is available in the target language, but an author-created button uses an inconsistent term for the same action.
Method. Identify which words are automatically supplied as system variables and which are authored content. Reuse supported system strings where their meaning is genuinely equivalent.
Verification. Run the action through screen readers and with the visible text. Consistency should help learners predict actions, not force the wrong word into a different context. Separate the format test from the education test. A file can be valid XML or XLIFF while a sentence misstates the rule. A rendered page can look polished while a branch chooses the wrong outcome. Use distinct tests for import integrity, meaning, interaction, accessibility and delivery rather than letting one green check mark stand for all five.
13. Establish a revision boundary
Situation. The English source changes while translators work on a frozen file, and the later import may be accurate only for an obsolete question.
Method. Record the approved source revision and classify subsequent changes by consequence: copy edit, changed meaning, altered assessment, or redesigned interaction.
Verification. Compare imported targets against the intended revision before publishing. An editor’s last-minute improvement should not silently invalidate the target-language sign-off. A learner encounters a sequence, not a list of phrases. They read a heading, interpret an instruction, make a choice and receive feedback. If one translated link or conditional message breaks that sequence, the remaining pages cannot compensate. Follow the task from discovery to completion and inspect the transitions.
14. Set translator and reviewer roles
Situation. The same person translates a complex safety section and approves their own interpretation without a separate content check.
Method. Separate linguistic review from subject approval for consequential material. dominKnow’s review workflow can route language editions to appropriate people, but role names alone do not guarantee expertise.
Verification. Ask who can certify the source rule and who can judge its target-language clarity. Both judgments must be represented where the learning outcome depends on exact instructions. Do not assume that content uniformity is always fair. People may need different examples or locally approved procedures to learn the same underlying competence. Preserve the shared objective while documenting legitimate variation; otherwise one supposedly universal edition may become incomprehensible or misleading to part of the audience.
15. Review AI-translated instructions
Situation. An automated translation renders a conditional safety rule elegantly but drops an exception and becomes materially misleading.
Method. Use AI output as a candidate, then inspect negation, quantities, legal or safety terminology, warnings and assessment feedback. Give high-risk segments greater scrutiny.
Verification. Create opposite test scenarios and ask a fluent reviewer which actions are allowed. A translation that yields the wrong action is not acceptable even if it sounds natural. Here is the deeper principle: language quality cannot be judged from isolated strings. The author must understand what the learner is trying to accomplish, which course mechanism presents the information and what evidence would establish that the learner understood it. If those three elements are disconnected, a fluent translation may still create a wrong or unfair learning experience.
16. Govern corporate terminology
Situation. A product name, process code or role title changes among language editions and confuses staff comparing the lesson with workplace systems.
Method. Maintain approved terms and explain when a label is official, translatable or accompanied by a local descriptor. Organization-level terminology preferences help reduce drift.
Verification. Check names inside menus, images, tests, transcripts and search labels. A learner should be able to locate the same real-world concept outside the course. Think about responsibility before efficiency. A translation service can propose wording quickly, yet it cannot know which workplace rule the source owner is empowered to change. Bring factual disputes back to that owner and ask a language specialist to explain the effect of the phrase. That keeps convenient automation from becoming accidental policy writing.
17. Localize numbers and thresholds
Situation. A threshold appears in several lessons and one target edition changes the decimal separator or unit without checking the original calculation.
Method. Distinguish language formatting from measurement changes. Preserve authoritative values and obtain expert approval for conversions or policy-specific local values.
Verification. Ask the reviewer to solve a sample problem using the displayed values. Verify that the quiz answer remains valid and that labels specify the units. An effective review has a defined acceptance question. Instead of asking only ‘Does this sound good?’, ask whether the learner can perform the specified decision without seeing the original course. A short scenario, followed by an explanation in the target language, often reveals more than several rounds of cosmetic correction.
18. Respect right-to-left direction
Situation. Arabic and English identifiers appear on the same page. The displayed wording is fluent but the arrow sequence and navigation order contradict it.
Method. Inspect directionality, icons, numbers and mixed-script identifiers in context. Avoid mirroring elements whose direction represents a real-world operation rather than reading order.
Verification. Test a branching interaction with a fluent right-to-left reader on a narrow screen. Verify that the intended next step is still visually unambiguous. Protect the future maintainer as well as today’s learner. A course may look excellent now and still be difficult to update if the project contains undocumented local exceptions, overwritten wording and ambiguous file names. Preserve the small decisions that are likely to matter when the original author is unavailable.
19. Translate visual information
Situation. An English diagram contains the only description of a critical decision, while the target edition offers only a translated paragraph beside it.
Method. Create a separate inventory for words embedded in images and for alternative text. Replace informative media where possible or provide a usable equivalent explanation.
Verification. Have a target-language learner explain the diagram without source-language knowledge. If the diagram cannot teach the expected concept, it needs more work. Separate the format test from the education test. A file can be valid XML or XLIFF while a sentence misstates the rule. A rendered page can look polished while a branch chooses the wrong outcome. Use distinct tests for import integrity, meaning, interaction, accessibility and delivery rather than letting one green check mark stand for all five.
20. Localize narrations and captions
Situation. The text changes but a video narration and caption track continue to use old terminology for the same procedure.
Method. Track audio, script and captions as linked assets with revision labels. Translate or replace them according to the delivery workflow and keep them synchronized.
Verification. Listen while watching the relevant interaction. Check technical names, numbers, pauses and the order of the warning relative to the action. A learner encounters a sequence, not a list of phrases. They read a heading, interpret an instruction, make a choice and receive feedback. If one translated link or conditional message breaks that sequence, the remaining pages cannot compensate. Follow the task from discovery to completion and inspect the transitions.
21. Preserve assessment discrimination
Situation. A quiz has one correct answer and three plausible distractors in English. In the target language two options become synonyms.
Method. Review the option set as a whole, preserving the intended knowledge test. Rewrite ambiguous distractors with subject-owner approval rather than treating each option in isolation.
Verification. Ask qualified target-language speakers to answer unaided and justify their choice. Compare the learning demand with the master question, not only the key letter. Do not assume that content uniformity is always fair. People may need different examples or locally approved procedures to learn the same underlying competence. Preserve the shared objective while documenting legitimate variation; otherwise one supposedly universal edition may become incomprehensible or misleading to part of the audience.
22. Keep feedback actionable
Situation. A translated feedback panel says ‘incorrect’ but does not explain the missed reasoning, while the source version provides a useful corrective hint.
Method. Translate the explanation and the next learning action, not just the correctness label. Review tone so errors teach rather than shame, and avoid vague pronouns.
Verification. Give a learner one deliberately wrong attempt and observe what they try next. Good feedback changes understanding, not merely the displayed color. Here is the deeper principle: language quality cannot be judged from isolated strings. The author must understand what the learner is trying to accomplish, which course mechanism presents the information and what evidence would establish that the learner understood it. If those three elements are disconnected, a fluent translation may still create a wrong or unfair learning experience.
23. Validate navigation and certificates
Situation. A learner finishes the course and sees a mix of localized screens and source-language end-screen or certificate labels.
Method. Inspect automatically translated system strings and custom-written course text separately. Do not rewrite official credential claims without approval.
Verification. Complete the course in each target language and inspect the final record, names, dates, labels and any completion claim that will be shared outside the course. Think about responsibility before efficiency. A translation service can propose wording quickly, yet it cannot know which workplace rule the source owner is empowered to change. Bring factual disputes back to that owner and ask a language specialist to explain the effect of the phrase. That keeps convenient automation from becoming accidental policy writing.
24. Test third-party links
Situation. An action link still points to a source-language help center with different region permissions, and the learner cannot perform the recommended step.
Method. Check destinations, redirects, language options and authentication needs. Explain when a related resource exists only in one language rather than claiming false coverage.
Verification. Open links from a test learner session in every supported region where practical. Verify they reach a usable destination and know what to do if access is restricted. An effective review has a defined acceptance question. Instead of asking only ‘Does this sound good?’, ask whether the learner can perform the specified decision without seeing the original course. A short scenario, followed by an explanation in the target language, often reveals more than several rounds of cosmetic correction.
25. Check published package selection
Situation. The editor displays three translations correctly, but the LMS launches an unexpected default language for some learners.
Method. Test published packaging, course-language selection, resume and completion in the target LMS. Treat any assumption about preference detection as a hypothesis to be verified.
Verification. Record a full learner path for each target locale. An authoring preview is not a substitute for evidence from the delivery environment. Protect the future maintainer as well as today’s learner. A course may look excellent now and still be difficult to update if the project contains undocumented local exceptions, overwritten wording and ambiguous file names. Preserve the small decisions that are likely to matter when the original author is unavailable.
26. Track changed phrases precisely
Situation. A reviewer corrects one mistranslated warning in the target course, but the following source update silently reintroduces the older translation.
Method. Keep a decision record tying the accepted wording to its source segment and approval. Flag protected local choices when regenerating translations.
Verification. Compare the release build to the corrected phrase during regression testing. The important property is persistent accuracy, not merely a successful one-time edit. Separate the format test from the education test. A file can be valid XML or XLIFF while a sentence misstates the rule. A rendered page can look polished while a branch chooses the wrong outcome. Use distinct tests for import integrity, meaning, interaction, accessibility and delivery rather than letting one green check mark stand for all five.
27. Plan a language-specific hotfix
Situation. A critical target-language mistake is discovered after publication, but other language editions remain accurate and in active use.
Method. Define a release process that can isolate the affected locale where the project architecture permits. Evaluate whether a broader package deployment is technically required.
Verification. Test the corrected phrase, related assessment and LMS continuation before replacing the published edition. Tell learners clearly when a material instruction has changed. A learner encounters a sequence, not a list of phrases. They read a heading, interpret an instruction, make a choice and receive feedback. If one translated link or conditional message breaks that sequence, the remaining pages cannot compensate. Follow the task from discovery to completion and inspect the transitions.
28. Document what an auditor needs to know
Situation. An administrator inherits a multilingual course and sees its status as ‘complete’ but cannot identify the approved source or regional reviewers.
Method. Record the source revision, enabled locales, chosen formats, review decisions, functional test outcome, known limitations and release identity.
Verification. Give another staff member the record and ask them to update one lesson. If they cannot determine which content is authoritative, the project is not yet maintainable. Do not assume that content uniformity is always fair. People may need different examples or locally approved procedures to learn the same underlying competence. Preserve the shared objective while documenting legitimate variation; otherwise one supposedly universal edition may become incomprehensible or misleading to part of the audience.
A four-week dominKnow translation plan
Week 1 — Establish governance and source truth
Identify the learning outcomes, most consequential decisions and the source revision to be translated. Determine whether the organization will publish one multilingual project or multiple language-specific projects. List supported target locales and who is able to approve each one. Do not fill every available language option merely because a service can generate text; an unsupported language is a maintenance promise without an owner.
Map course elements including modules, learning objects, pages, pop-ups, variables, links, glossary entries, assessment feedback and media. Separate stable common material from region-dependent facts. Choose a small pilot that includes rich text, a variable, a branch and an assessment so the translation workflow is tested on meaningful complexity.
Week 2 — Exchange and review context-rich language
Select the appropriate translation route: AI assistance or manual exchange using Excel, XML or a compatible XLIFF format. Explain the course purpose, audience and actions to the language reviewer. If exporting structured files, protect identifiers, CDATA, inline markup and variable tokens. Translate the meaning of an instruction rather than treating the string as a decorative label.
Use the review and approval process to distinguish language polish from factual disputes. Keep a decision record for high-impact terms and preserve any accepted target-language changes. Inspect the imported content in the editor, because a correct file import does not guarantee that the words work in their actual interaction.
Week 3 — Prove the learner journey works
Ask a fluent learner to complete one representative activity without source-language assistance. Include an incorrect answer, a retry, a pop-up and a link. Check whether the feedback explains the intended idea and whether assessment distractors remain meaningfully distinct. Inspect target-language headings, alt descriptions, keyboard access and screen-reader output where relevant.
Test the published course or package in its destination environment. Observe language selection, launch, resume, completion and any scoring records. If a target course includes local procedures, confirm those instructions with the proper regional owner. Record what was tested and which issues remain, rather than using a vague ‘translation approved’ status.
Week 4 — Release and maintain approved editions
Approve the course by competence, language, interaction and delivery. Prepare a short release note identifying the master, target versions, translation method, validated branch paths and accepted limitations. Stop if a critical warning or scored item is ambiguous. It is better to postpone an edition than to publish a learner-facing instruction that the responsible reviewer cannot explain.
Retain the approved terminology and export revision, record reviewer decisions and rehearse a small future update. When new source material arrives, identify affected target segments rather than recreating every language blindly. A maintainable multilingual system is one in which the next editor can make a controlled change without undoing the last team’s carefully reviewed work.
Frequently asked questions
What is the difference between one multilingual project and separate projects?
A multilingual project groups several language versions within one course structure and can simplify common maintenance. Separate projects can offer independent content and governance for each edition. Compare these models against regional requirements, the nature of your assessments, language selection, reviewer availability and LMS deployment before choosing.
Does dominKnow ONE provide AI course translation?
Its official localization materials describe an AI Translator and workflows for generating and reviewing translated courses. Availability can depend on deployment and account settings. Treat generated text as a proposed translation, then confirm high-risk rules, interaction labels, subject terminology and graded answers before publication.
Which translation exchange formats are supported?
dominKnow’s product and community documentation describe Excel, XML and XLIFF exchange options, including XLIFF 1.2 and 2.1 in language configurations. Specific workflows vary; run a representative round-trip test with your translation environment to verify that its particular version and markup survive import.
Why do translated Excel exports include context columns?
Course, module, learning-object, page and element information helps the translator understand where each string is shown. A short word such as ‘Hold’ can represent a button, a noun or an instruction depending on that context. Reviewers should use the contextual information instead of treating the file as an isolated vocabulary list.
Do all built-in interface labels require translation?
Not necessarily. dominKnow documents automatically supplied translated player strings and related system variables. Authors still need to inspect custom labels and check whether the built-in wording conveys the correct action for their course. Reuse is useful only when the meanings align.
Can reviewers approve an automatically translated safety module without testing it?
For consequential learning, automatic fluency alone is weak evidence. A reviewer should test conditional rules, negation, numeric thresholds, decisions and corrective feedback. Ask a target-language learner to demonstrate the intended behavior in a realistic scenario before signing off.
How do you protect variables and markup?
Explain to translators which tokens, tags and file boundaries are structural. Use supported protection features in translation tools where possible, then compare imports and test dynamic content with several values. A placeholder can survive visually while still producing an ungrammatical or misleading sentence.
What should be recorded before release?
Keep the approved source revision, supported locales, translation route, reviewers, unresolved limitations, assessment checks and delivery tests. A concise change record makes the course auditable and easier to maintain when the same topics or policies change.
Can language settings replace a local style guide?
No. Enabling a language is an operational setting. A style guide tells translators what voice, register, terminology, unit conventions and audience assumptions are acceptable. Both are needed if the team wants predictable courses rather than a collection of individually plausible translations.
How should multilingual learning results be compared?
Compare like-for-like objectives and learner populations, then inspect item-level evidence rather than assuming score differences are caused by translation. Confirm that LMS completion data belongs to the correct course edition. Results can flag a problem, but a targeted comprehension test is needed to establish its cause.
Primary sources and related guides
The dominKnow ONE translation and localization overview, administrator language-configuration guidance and translation export-file instructions describe the product’s current features and practical file handling. For the broader education-translation framework see the Master Art of Translation: Education and Learning Systems. Compare the adjacent Easygenerator multilingual courses guide for a different authoring approach.
The professional result is not simply a translated file. It is a course whose approved facts remain true, whose interactions continue to teach, whose learners can understand feedback and whose release can be inspected and maintained. Make every language edition earn that confidence through meaning, behavior and evidence.