Translate ISO 26262 safety documents across languages by preserving the distinction between ASIL A–D, QM, hazard analysis and risk assessment (HARA), safety goals, and safety requirements. A technically fluent translation is not a list of bilingual equivalents. It is a defensible transfer of the engineering relationships that allow another team to identify which hazardous event, assumption, classification and work product a sentence concerns.
When an automotive safety case crosses a language boundary, one incorrect noun can redirect the reader toward the wrong engineering decision. ‘Exposure’ is not the probability that a chip fails; ‘controllability’ is not feedback-loop gain; an ASIL is not a consumer car-safety star rating. These distinctions matter for translators, multilingual engineering teams, component suppliers, independent assessors and anyone reviewing safety-related electrical or electronic systems.
This guide explains how to translate ISO 26262 vocabulary, HARA tables, risk classifications, hardware-safety metrics, development gates and evidence claims without inflating compliance or weakening requirements. It focuses on one practical search intent: how to read and translate an automotive functional-safety document accurately while leaving the controlling technical decisions with the qualified engineering owner.
One-minute reading map
Translation task: carry a controlled automotive functional-safety statement into another language without changing its risk classification, requirement strength, trace links or evidence scope. Core distinction: ISO 26262 is a framework for functional safety of safety-related road-vehicle electrical/electronic systems; translating its records does not certify a product. Working rule: read each sentence beside the source identifier, linked hazard, classification, responsible owner and applicable standard edition.
What the document is actually doing
Did you know that a functional-safety requirement can be linguistically simple yet technically difficult to translate? Consider a sentence that directs a controller to detect a fault and move to a defined safe state within a stated interval. A literal language substitution may preserve every dictionary word but lose whether detection, reaction or transition is timed. Engineering readers could then implement different behavior. Professional translation must keep the original temporal logic, trigger, subject, action and acceptance criteria together.
ISO 26262 uses a lifecycle approach rather than a single pass/fail label. The item is defined, hazardous events are assessed, safety goals and requirements are derived, and evidence is collected across development and operation. A translated line should disclose whether it is an assumption, requirement, design decision, verification result, validation conclusion or residual issue. These types of sentence are not interchangeable. When the target language lacks a neat noun for one of them, an accurate explanatory phrase is preferable to a misleading short label.
The most useful translation unit is not always the individual sentence. It is often a source sentence plus its row identifier, abbreviation, referenced figure and governing table. Readers should not have to guess whether ‘control’ means a driver avoiding harm, a system acting on an actuator, or an audit control. Context separates these interpretations. That is why the article focuses on specific reader decisions rather than producing a competing all-purpose translation handbook.
What the categories do—and do not—mean
Severity concerns possible injury, exposure concerns the operational situation, and controllability concerns the ability to avoid harm in the hazardous event. Those categories contribute to the classification outcome through the applicable method; they are not three interchangeable percentages. ASIL D represents the highest integrity demands within ASIL A–D. QM is not a marketing mark. A warning in a safety analysis may be conservative, but a translator is not authorized to weaken it because it sounds alarming in the destination language.
The terminology is particularly sensitive to implied certainty. ‘A safety mechanism is intended to detect specified faults’ is not equivalent to ‘the safety mechanism detects every fault’. ‘Assessment completed with findings’ is not equivalent to ‘fully safe’. ‘ASIL D safety goal’ is not equivalent to ‘ASIL D vehicle’. Readers may make procurement, test and architecture decisions based on precisely these distinctions. Translation therefore includes a check against both overstatement and understatement.
A better reading habit for multilingual teams
Begin with the noun that is being classified. Next identify who performs an action, what triggers it, and when it must occur. Then find the supporting work product and any stated uncertainty. Finally read the line against the practical consequence of getting it wrong. This small sequence catches many errors that ordinary bilingual proofreading cannot. Language quality and engineering validity remain separate approvals, but a well-designed translation process makes them cooperate rather than compete.
Thirty-two translation situations that change engineering meaning
1. ASIL A, B, C and D
Source situation. The source table grades a safety goal ASIL D, not the vehicle itself. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Calling D a universal crash-test grade, or treating A as a general quality award, misrepresents the safety integrity classification. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Keep the letter attached to the named safety goal or inherited safety requirement; retain the assessment context, standard edition and the underlying hazardous event. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Compare the translated table row with the HARA record and ask which goal actually carries the classification. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
2. QM versus an ASIL
Source situation. A scenario is recorded as QM after the hazard analysis rather than assigned ASIL A. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Expanding QM as an assurance that no safety issue can occur creates an unsupported conclusion; it means the ASIL classification did not require the ASIL measures for that item under that analysis. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Preserve QM as a classification and distinguish it from the quality-management process that remains applicable. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check the HARA rationale, the version of the item definition and any later reassessment. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
3. Hazard and hazardous event
Source situation. A hazard describes a potential source of harm arising from malfunctioning behaviour; the hazardous event adds the operational situation. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Translating both terms as an incident that already happened collapses prospective risk identification into a retrospective accident report. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Render the potential condition and the particular driving situation separately, including operational context. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Confirm whether the source is a hazard catalogue, a scenario row or an actual occurrence record. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
4. HARA
Source situation. A document heading says Hazard Analysis and Risk Assessment and cites a particular item revision. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Treating HARA as a generic roadside inspection hides its structured purpose and may obscure revision dependencies. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Expand the acronym once, preserve HARA thereafter and tie its results to the item and defined operating scenarios. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Inspect the document header, controlled revision and decision authority before updating a translation memory. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
5. Item definition
Source situation. The item definition specifies the function, interfaces, boundaries, assumptions and environment for a driver-assistance feature. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Rendering item as a physical part number only can narrow the assessed boundary and exclude relevant interfaces. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Describe the functional system under analysis rather than a random inventory unit. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Trace each translated system boundary to the source diagrams and list assumptions that were explicitly excluded. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
6. Severity S
Source situation. The HARA row labels severity with an S category based on potential injuries. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Converting an S rating into a probability of occurrence changes its dimension; severity is about possible harm, not how often the drive happens. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Keep the coded S category unchanged and translate its contextual explanation with its original qualifications. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check the standard’s permitted category interpretation and the injuries assumed in the scenario. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
7. Exposure E
Source situation. The document assigns an E category describing frequency or duration of exposure to the relevant operational situation. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. A mistranslation as component failure probability could make the evidence look stronger or weaker than it is. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Explain exposure to the situation, not failure rate, while preserving any source-specific exposure rationale. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check that the road, weather, traffic or operating condition used in the row survives translation. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
8. Controllability C
Source situation. The C rating represents the ability of people involved to avoid harm under the hazardous event. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Translating controllability as software control-loop stability erases the human avoidance aspect of the classification. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. State who is assumed to respond, the kind of response and any limits recorded by the analysis. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Review driver assumptions and whether vulnerable road users also matter to the scenario. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
9. S-E-C combination
Source situation. A matrix combines severity, exposure and controllability to assign an ASIL or QM result. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Adding scores arithmetically or inventing a numeric ASIL conversion is not an acceptable localization shortcut. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Preserve the original categorical triplet and the published lookup result without recomputing it from prose. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Compare the row against the exact source matrix; query discrepancies instead of silently fixing them. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
10. Safety goal
Source situation. The HARA produces a safety goal, such as preventing unintended torque under a stated context. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Turning a goal into a claim of zero failures or a test result promises more than the document says. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Keep the normative intention distinct from implementation evidence or validation outcome. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Verify the goal ID, ASIL allocation, linked hazard and exact scope of the requirement. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
11. Safe state
Source situation. An engineering requirement describes a specified safe state for a particular fault or operating condition. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Replacing it with the vague phrase ‘turn everything off’ can be unsafe where a controlled transition is required. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Translate the defined state and its entry criteria; preserve the source’s transition logic and limitations. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Ask engineering reviewers to confirm any term for controlled shutdown, degraded operation or torque limitation. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
12. Fault-tolerant time interval
Source situation. A requirement specifies an interval in which a fault must be controlled before a hazardous event can occur. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Confusing the interval with total service life or average repair time destroys its timing implication. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Carry the acronym and units forward, including the defined start and stop conditions and the safety mechanism it constrains. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Reconcile the interval with diagnostic and fault-reaction timing tables. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
13. Failure versus fault
Source situation. The report distinguishes an observed functional failure, an internal fault and the resulting error state. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Using one target-language word for all three may make diagnostic sequences and root-cause arguments impossible to reconstruct. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Maintain separate approved equivalents and annotate ambiguous usage instead of smoothing all distinctions away. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Build a local glossary with definitions and counterexamples from the project. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
14. Functional safety concept
Source situation. The functional safety concept allocates functional safety requirements derived from safety goals. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Mistaking it for the detailed hardware circuit design skips a critical abstraction level. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Translate the concept as a structured requirements allocation, not merely a marketing philosophy. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Compare the required outputs and trace links with the technical safety concept. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
15. Technical safety concept
Source situation. The technical safety concept defines how technical safety requirements meet higher-level functional requirements. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Rendering technical as only a software specification can misplace hardware and interface obligations. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Keep system architecture, hardware, software and interfaces in view whenever the source includes them. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Follow traceability links between functional and technical requirements. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
16. FSR and TSR identifiers
Source situation. Rows use FSR-012 and TSR-044 as controlled requirement identifiers. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Translating or renumbering identifiers breaks cross-reference integrity even if the prose sounds natural. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Preserve IDs byte for byte while translating explanatory text and adding target-language labels only where permitted. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Run an automated comparison of requirement identifiers across versions. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
17. ASIL decomposition
Source situation. The architecture discusses ASIL decomposition and allocated requirements with independence assumptions. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Translating decomposition as automatically reducing every component’s required integrity omits constraints and evidence obligations. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. State which requirements are decomposed and preserve any notation, assumptions and independence conditions. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Escalate technical rewording to the safety engineer; do not invent a decomposition rule. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
18. Freedom from interference
Source situation. A design claim concerns separation between elements of differing ASIL contexts. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. A simple translation as ‘no interference from users’ changes the system-safety issue into a user-interface statement. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Describe interference in the engineering system and retain referenced mechanisms, interfaces and partitions. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Compare memory, timing, communication and resource interference contexts recorded in the source. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
19. Dependent failures
Source situation. The analysis considers common cause and cascading dependencies, not just independent component failures. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Calling every dependency a simultaneous independent failure can erase the causal link the analysis exists to assess. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Preserve source distinctions among dependent, cascading and common-cause concerns when present. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Follow the diagrams and list of shared resources before approving terminology. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
20. SPFM
Source situation. A hardware metric appears as SPFM with a percentage and a project acceptance target. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Expanding SPFM incorrectly or translating its percentage as an actual field failure rate changes the measurement. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Use the approved expansion single-point fault metric, retain numerator conventions as defined in the applicable method and never infer a fleet rate. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Match the metric calculation source, scope and supporting FMEDA. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
21. LFM
Source situation. The report lists an LFM percentage alongside diagnostic coverage information. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Equating latent fault metric with maintenance-downtime proportion is a category error. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Keep latent-fault terminology separate from repair performance and reporting dates. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check that the reported metric refers to the safety-related hardware architecture assumed by the analysis. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
22. PMHF
Source situation. The safety analysis refers to probabilistic metric for random hardware failures in a specified safety goal context. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Treating PMHF as the probability of any crash per hour would be both exaggerated and unsupported. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Retain the metric name, numerical notation, exponent, units and claimed scope without broadening the interpretation. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Recheck all powers of ten and compare the source’s exposure assumptions. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
23. Diagnostic coverage
Source situation. The source documents a test and a claimed percentage of detected faults. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. A high percentage must not become ‘all faults detected’ in the target language. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Translate what failure modes and mechanisms were included, and do not overstate the coverage basis. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Review any listed excluded faults and measurement or simulation conditions. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
24. FMEDA and FMEA
Source situation. The file distinguishes failure modes and effects analysis from failure modes, effects and diagnostic analysis. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Flattening the two acronyms conceals whether diagnostic analysis was actually performed. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Retain each acronym and give each its correct approved expansion; do not imply a FMEDA exists because an FMEA exists. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check cited annexes and calculation worksheets before referring to coverage numbers. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
25. Safety mechanism
Source situation. A watchdog, plausibility check or redundant sensor is described as a safety mechanism. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Calling a feature a safety certification or guaranteeing complete protection misreads an engineering control as an outcome. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Keep the actual function and the fault it addresses visible in the translated sentence. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check detection, diagnostic and response descriptions as a connected chain. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
26. Confirmation review
Source situation. A project plan distinguishes confirmation reviews, functional safety audits and functional safety assessments. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Translating all of these as an informal signoff can erase different objectives and degrees of independence. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Preserve the named activity, independence requirement where relevant, owner and recorded conclusion. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Reconcile the translated deliverable register with the source work-product list. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
27. Safety case
Source situation. The argument links claims, evidence and reasoning in a structured safety case. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Translating it as a single certificate or proof that an accident cannot occur promises a result never established. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Explain that the case assembles an argument supported by evidence within a defined scope. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check claim IDs, evidence references, limitations and open findings. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
28. Change impact analysis
Source situation. A supplier updates a component revision and the safety engineer evaluates changes to safety work products. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Treating minor in commercial terms as automatically minor for functional safety can bypass re-evaluation. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Preserve the explicit impact decision, unresolved questions and affected work-product IDs. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Match the change ticket with the assessed baseline and the issued translation version. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
29. ISO 26262 versus SOTIF
Source situation. A project distinguishes malfunction-driven functional safety from hazards associated with intended functionality and performance limitations. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Replacing ISO 21448 SOTIF with a synonym for ISO 26262 misstates the standards scope. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Keep the standards separate and explain their differing questions only to the extent covered in the source. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check that conclusions are not carried from one assessment into the other by translation alone. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
30. Release versus approval
Source situation. The development gate records release for the next phase while some issues remain open. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Translating release as full regulatory approval or proof of safety converts project governance into a stronger claim. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Use the source’s actual gate disposition, named signatories and limitations. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Compare conditional approvals, waivers and closure dates without normalizing them. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
31. Safety requirement ‘shall’
Source situation. A controlled requirement uses ‘shall’, while a rationale or explanatory note uses ‘should’ or ‘may’. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Turning mandatory language into advice changes contract-like engineering obligations; reversing the direction can impose unapproved requirements. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Apply a modals glossary consistently and preserve distinctions among mandatory, recommended and permitted statements. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Diff every translated requirement for shall, should, may, must, must not and conditional wording. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
32. OEM and supplier responsibilities
Source situation. An OEM allocates responsibilities for item definition, safety goals and supplier evidence. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. A translated sentence that assigns full responsibility to a supplier when the source retains integration duties changes governance. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Identify the owner of each safety activity and preserve handoff and acceptance conditions. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Validate RACI entries, interface agreements and sign-off records. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
33. Versioning and baselines
Source situation. A source cites ISO 26262:2018 and a released project baseline; a draft revision may also be circulating. A reader who has only the target-language text must still be able to identify the original decision and its object. The code, category or technical term should not become a free-standing slogan; it should stay attached to the same sentence, controlled item and documentation purpose.
The consequential mistranslation. Describing a draft as an already adopted requirement or silently updating a 2018 reference is misinformation. Here a change that might look minor to a copyeditor could modify the assumptions on which the system was assessed. A polished sentence is not a successful translation if it leads the technical reader to a decision the original never supported.
How to render the meaning. Keep the edition, status and effective dates exactly as documented; link proposed changes separately. In practice, keep identifiers and technical symbols as stable anchors, then make the relationship intelligible in natural target-language prose. Explain an acronym the first time it appears when appropriate, without inventing a replacement code or an unauthorised compliance statement.
Reviewer test. Check official standard status at publication and annotate any uncertainty in the reference note. A second reviewer should be able to answer what was asserted, what evidence would verify it and which controlled version was used. If those answers differ between the source and translated versions, the passage requires a documented terminology query rather than a silent rewrite.
Ten-step quality assurance workflow
Step 1: Locate the controlled item definition
Record system function, boundaries, interfaces and operational assumptions first. A translation can be grammatically flawless yet address a different system if the reference version changed. Tag all source documents with revision and owner. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 2: Build a safety-specific terminology ledger
Create stable target-language equivalents for hazard, hazardous event, safety goal, fault, failure, error, exposure, controllability, safety mechanism, functional safety concept and technical safety concept. Note cases where one familiar word is misleading. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 3: Separate numbers, categories and claims
Keep ASIL letters, S/E/C categories, percentages, powers of ten, timing intervals and requirement identifiers unchanged unless the authorized source itself changes. Every unit conversion needs explicit dimension checking. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 4: Map the hierarchy
Trace source item definition to HARA row, then safety goal, functional requirement, technical requirement, architecture and verification. Do not imply a new causal relationship because it reads more smoothly. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 5: Protect modal verbs
Review each shall, should, may, need not and must not separately. The strength of obligation is an engineering datum, not a stylistic preference. Translate the conditional trigger and the action in the same sentence. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 6: Preserve scope and exclusions
Carry known limits, safety assumptions, intended functions, excluded functions, vehicle classes, operating scenarios, safety cases and unresolved issues into the translation. Avoid a broad generic claim of compliance. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 7: Validate acronym expansions
Check local standard edition and project conventions before expanding HARA, FSR, TSR, PMHF, LFM, SPFM, FMEDA or DFA. Do not substitute a common-looking expansion for a documented one. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 8: Use paired reviews
Have a trained language reviewer assess readability and a qualified safety-domain reviewer assess technical preservation. Neither role alone is sufficient for a safety-critical translation workflow. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 9: Test hostile interpretations
Ask what a hurried supplier or tester might do differently after reading the translation. If an interpretation could change a requirement, reroute a test or hide a limitation, revise or flag the sentence. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Step 10: Create a return trail
Store source revision, translated revision, glossary changes, reviewer comments and acceptance decisions. A later update should be able to identify exactly which output text came from which controlled source paragraph. Record the reviewer and the outcome alongside the source segment. The deliverable is not only fluent text: it is a traceable rendering that allows another reader to reconstruct the original controlled meaning.
Frequently asked questions
Is ASIL D a score for the whole vehicle?
No. In ISO 26262 work products, ASIL is associated with the hazardous-event analysis and safety goals, and flows into safety requirements according to the applicable process. Avoid attaching one classification to every vehicle function. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Does QM mean the component cannot cause harm?
No. QM is a classification in the applicable analysis, not a zero-risk claim. The original item assumptions and quality obligations still matter. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Can the translator calculate an ASIL from words alone?
No. The categorized inputs, source matrix and engineering judgement belong to an authorized HARA process. Translation should preserve, not invent or silently recompute, that determination. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Should ASIL names be translated?
Explain them in the target language when useful but keep the recognizable ASIL identifier and letter unchanged. Avoid coining a locally appealing name that suggests a different certification scheme. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Are FMEA and FMEDA synonyms?
No. A source may use both; their analytical content differs. Glossary entries should preserve the exact named analysis and the data it actually contains. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Why not replace shall with must everywhere?
Some style guides do, but the decision must be controlled. The key is preserving obligation consistently, including prohibitions, conditions and scope, rather than making ad hoc stylistic changes. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Does translated approval mean regulatory approval?
Not automatically. The source may refer only to an internal release gate, a conditional sign-off or a particular assessment. Keep the kind of approval and its authority specific. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
How should we handle an uncertain technical term?
Log a terminology query with the source extract, proposed equivalents, the suspected ambiguity and its possible effect. Leave the code and controlled references untouched until resolved. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
What if the source contradicts the referenced table?
Do not repair the conflict silently. Preserve the source value, identify the discrepancy and escalate to the controlled document owner for correction. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Who verifies translated safety metrics?
A qualified engineer or authorized technical reviewer should confirm the interpretation, scope, calculations and applicability. A translator can verify faithful transcription but should not independently certify design safety. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Does this guide replace the ISO standard?
No. Consult the applicable authorized edition, contractual requirements and project processes. This educational guide supports reading and translation quality, not safety engineering approval. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
How does this article connect to wider language learning?
It applies a general language principle to a bounded technical problem: words carry relationships and obligations. The wider translation and vocabulary learning architecture should remain the canonical place for broad methods. Where a project-specific definition differs, follow the controlled source and flag the difference explicitly rather than blending incompatible conventions.
Worked review: from an English HARA row to a second language
Imagine a row whose situation is ‘unintended acceleration during low-speed manoeuvring’, whose severity, exposure and controllability categories are recorded, and whose safety goal refers to preventing an unwanted torque request. A translator begins by identifying what is hypothetical, what is categorical and what is normative. The event is prospective. Its S/E/C ratings belong to that row. The safety goal specifies an objective, not proof of implementation. A proposed translation that calls the event an accident already suffered would change the timeline. One that says the torque control is guaranteed to work would transform the goal into an outcome. Both fail the review even if they sound confident.
Now suppose the technical safety requirement states that a diagnosed signal discrepancy shall result in a controlled transition within a defined interval under certain conditions. The trigger, the response, the time unit and the conditions are four distinct elements. In a good translation each survives unchanged. The reviewer checks the requirement ID, the referenced timing diagram, the matching test procedure and the governing baseline. When the target language prefers a longer explanatory structure, clarity is welcome provided no extra exception, guarantee or response is introduced.
What to publish and what to keep private
Public educational content can explain distinctions, hypothetical scenarios and reading techniques. It should not disclose confidential HARA rows, supplier interfaces, unpublished safety cases or internal diagnostic strategies without authorization. Even a technically sound translation may be unsuitable for public release if it reveals sensitive implementation details. A publishing review therefore checks audience, confidentiality, the rights to standards content and the status of the underlying source. Proprietary standards text should be referenced rather than reproduced.
Sources and limits
Primary source: ISO’s ISO 26262 Part 9 standard project page (the displayed draft status must not be mistaken for adoption). For functional-safety context, see TÜV SÜD’s overview of automotive ISO 26262 and TÜV Rheinland’s explanation of automotive functional safety. Verify the controlling edition and the applicable contractual documents before making an engineering decision. This article is educational interpretation, not certification or a substitute for a licensed standard.
For broader language concepts, return to the eduKateSG learning library; the master translation architecture remains the owner of general methods. This page is a specialist worked example about ISO 26262 documents, not a replacement translation or vocabulary hub.