Translate | ISO 3037:2022 Corrugated Edge Crush Strength, ECT, Flute Direction and Packaging Test Reports

How do you translate an ISO 3037:2022 edge crush test report for corrugated cardboard without confusing ECT values, flute direction and carton stacking strength? The search intent is practical: a manufacturer, laboratory, importer or packaging buyer needs an accurate English or multilingual rendering of edgewise crush resistance that preserves the measured property, units, sampling identity and laboratory limitations.

ISO 3037:2022 covers a non-waxed edge method for determining edgewise crush resistance of corrugated fibreboard. Its public scope includes single-, double- and triple-wall board, with loading along the flute axis, and notes conditions concerning buckling or tilting. This guide explains the language and evidence needed to communicate results faithfully; it is not a substitute for the licensed test procedure.

If you are translating ECT kN/m figures, carton compression laboratory reports, supplier declarations or corrugated board specifications, begin with the distinction between a specimen’s edgewise resistance and the behaviour of a whole finished box. That single distinction prevents many attractive but inaccurate translations. We will then move through test identity, sample preparation, numerical reporting, worked examples and sign-off.

This is an original educational interpretation of technical-document language. All numerical scenarios are fictional illustrations, not values from the standard, accredited test results or suggested product acceptance thresholds. Apply the controlled specification, laboratory report and qualified engineering review to real purchases.

What the certificate can—and cannot—prove

A technical translation answers a narrow question accurately. It does not issue a new certificate, redesign the carton or convert a material test into a universal promise about delivery. The difference becomes easy to see once the report is read field by field. Use the following guide as a sequential evidence audit, not as a replacement for the issuing laboratory’s judgement.

1. Understand which strength is actually measured

The expression edgewise crush resistance belongs to the strength of corrugated fibreboard when a prepared piece is compressed on its edge. It is not a claim that an assembled box can survive a particular stack height. Translate the material-level property first, retaining its test reference, orientation, units and specimen definition. A buyer who sees merely ‘box strength’ may assume that an entire finished package has been qualified.

Imagine a multilingual purchasing specification saying ‘ECT 5.4 kN/m’ beside a box-compression requirement expressed in newtons. Both values can concern shipping performance, yet they describe different test objects. Calling the two values equivalent erases the distinction between a fibreboard sample and a complete box. If the supplier changes flute, board construction or dimensions, one result cannot simply stand in for the other.

A careful target-language sentence might read, ‘Edgewise crush resistance of the corrugated board, measured by the cited method: 5.4 kN/m.’ The translation should then preserve the separate box-compression result on its own line. Ask which material, which finished package, and which test generated each number before deciding that two reports contradict one another.

2. Read the standard identifier as part of the evidence

ISO 3037:2022 identifies a particular published edition for determining the edgewise crush resistance of corrugated fibreboard by a non-waxed edge method. Edition control is not a decorative footer. Reports can refer to previous procedures, internal adaptations or different national documents, and a translator should not silently upgrade an older test citation merely because a newer edition exists.

If a certificate references ISO 3037:2007 and a product brief references ISO 3037:2022, the bilingual output must preserve both citations as found. Engineering or quality personnel can decide whether one must be reissued. Replacing the older date with the latest publication would manufacture a degree of compliance that the source certificate did not establish.

Create a small source register recording document title, issuing laboratory, document number, revision and cited test edition. Align every translated value with that register. The reader should be able to trace a target-language assertion back to a specific line of an original document without relying on the translator’s memory.

3. Keep the non-waxed edge method distinct

The published ISO description specifies a non-waxed edge method. This matters because a testing procedure is not identified merely by the quantity it claims to measure. Specimen preparation and edge support can affect the meaning of a result. A translated instruction that introduces waxed edges or substitutes another procedure changes the method rather than clarifying it.

A client may abbreviate the source as ‘edge test’, which is intelligible in conversation but insufficient in a contractual comparison. Translate such informal language in a way that remains faithful to its level of precision, then request the actual standard reference where decisions depend upon it. Do not expand the abbreviation into a certified method without evidence.

In a technical annex, preserve ‘non-waxed edge’ or its precise approved target-language equivalent alongside ISO 3037:2022. Where procurement uses a familiar local expression, place that wording in an explanatory gloss rather than replacing the method identity. The test method is part of the result’s identity.

4. Follow the direction of the corrugation flutes

The standard’s public description says that compressive force is applied in the direction of the flute axis. A flute is the formed wave in the corrugated medium, not simply a printed line on a carton. In a drawing or test photo, orientation can be the difference between an understandable laboratory setup and a misleading diagram.

Consider a translator given a photograph rotated ninety degrees for a catalogue layout. The caption ‘vertical flutes’ now appears beside an image in which the board has been rotated. A reader may assume the test load acts in a direction it did not. Translate the description against the original specimen orientation rather than relying on page design.

The target report should preserve phrases such as ‘parallel to the flute axis’ where they genuinely appear. If the source shows arrows, redraw or place them only under an authorized technical review. Translating a diagram requires preserving relations, not merely moving labels into another language.

5. Distinguish flute profile from the flute axis

Single-wall, double-wall and triple-wall board can contain one or more corrugated mediums. Letters used for flute profiles may describe geometry or grade conventions, but they do not themselves tell an inspector every condition of a particular specimen. The word ‘flute’ also appears in musical contexts; a literal multilingual dictionary entry is especially unhelpful here.

A purchasing email says ‘B flute, 3-ply’ while a laboratory report says only ‘corrugated board’. A translator who appends the purchasing description to the laboratory result makes the test sound more specifically documented than it is. These records may refer to the same product, but that link must be established by product code and batch evidence.

Keep the profile, wall construction, ply notation and tested specimen identity in separate fields. If a product code is missing, mark the relationship unresolved and ask the originating organization. An accurate translation can state exactly what a source does not identify, and that honesty often prevents the most expensive packaging mistakes.

6. Understand single-, double- and triple-wall scope

ISO’s published scope includes single-wall, double-wall and triple-wall corrugated fibreboard. These terms identify configurations of board construction, not simply the thickness of an arbitrary paper sheet. Translating ‘double-wall’ as ‘double thickness’ can suggest a geometric relationship that the manufacturer has not established.

Suppose a buyer compares two boards, both described informally as ‘two-layer’. One is double-wall corrugated board, while the other is a single-wall sheet with a different liner combination. The source-language shortcut may conceal a construction difference. A translated technical declaration should use the specific board-construction term supported by the specimen description.

Check cross-section sketches, supplier datasheets and internal material codes. Where the language is ambiguous, show the ambiguity as a query rather than inventing a structure. No translation can compensate for an incorrectly matched physical sample.

7. Separate test specimens from whole boxes

The standard permits specimens taken from corrugated boxes and other converted products within its scope. A specimen cut from a finished box is still a material sample when tested. It does not automatically reproduce the load case of a complete box with folds, joints, tape, openings and the actual contents inside.

A shipment dispute may contain photographs of crushed cartons and an unrelated edge-crush laboratory certificate. One language might call both ‘compression strength’. A useful translation preserves the certificate accurately while describing the field damage as an observation. It does not use the certificate as proof that no box failure could have occurred.

Translate ‘sample taken from converted box’ as sampling history, not as ‘box tested under service conditions’. Reviewers should be able to reconstruct where the sample came from and which object the result belongs to. That distinction is essential when a purchaser asks why an ECT number did not predict every transport outcome.

8. Preserve specimen identity and traceability

Names such as sample A, lot 24-16, reel 8 and board code BC-220 are identifiers, not descriptive prose to polish. If a translator changes a code for readability or merges two similar labels, the language version can no longer be matched unambiguously to laboratory photographs and retained specimens.

A laboratory may receive several grades of board from the same client on the same day. An attractive translation template could repeat the same generic ‘corrugated sample’ heading for all of them. The result looks consistent but loses material identity. Never let uniform layout override source traceability.

Build a bilingual index with customer identifier, manufacturer code, lot, sampling location, date, sheet or sample number and report revision. Keep the identifiers byte-for-byte where possible, while explaining the field names. Ask another reviewer to find the physical source sample using only the translated report.

9. Do not mistake peak load for normalized edge strength

Compressive force may be recorded in newtons, while a reported edgewise resistance may be normalized by specimen dimension and expressed per unit length, often as kN/m. A number without the unit and normalization context is not safely translatable. Even when arithmetic is straightforward, a test report must govern the definition of the denominator.

For a purely illustrative calculation, a maximum force of 480 N divided by a supported specimen length of 0.10 m gives 4,800 N/m, or 4.8 kN/m. This arithmetic exercise is not a prescribed ISO specimen dimension, correction rule or pass criterion. It demonstrates why 480 N cannot be relabelled as 480 kN/m.

Carry numerical values, units and source precision together. A reviewer should check whether the report contains raw instrument force, a calculated result, an average or an acceptance threshold. The translator’s role is to preserve which kind of number it is, not to recompute or reinterpret the laboratory outcome.

10. Keep N, kN and kN/m separate

The uppercase N denotes newtons, kN denotes thousands of newtons, and kN/m includes normalization per metre. In a multilingual spreadsheet, a slash may vanish in a font substitution or a cell may be formatted with only ‘kN’. Those small typographic changes can alter the apparent physical property significantly.

Imagine a product brochure showing ‘4.9 kN/m’ but an imported table rendering the value as ‘4.9 kN’. A customer may think the figure describes the compressive capacity of a complete carton. Translate field headings with the unit integrated, verify superscripts and symbols after export, and compare the final PDF to the source.

Never infer a unit from a neighbouring row if the source field is empty. Query missing information and distinguish a confirmed source unit from a translator-supplied explanation. Mechanical reproduction of numbers without their dimensional meaning is not an adequate translation of engineering evidence.

11. Understand the end-point and invalid deformation

The ISO scope describes application where buckling or tilting does not occur during measurement. That caveat matters because an apparent maximum load might be influenced by the way a specimen deforms rather than by the edgewise compression behaviour the method intends to observe. Translators should not call every test-run value valid merely because an instrument produced a graph.

A fictional laboratory note says ‘specimen leaned; result excluded’. A careless translation shortens it to ‘specimen failed’, and the customer reads the entry as a product failure. The original may instead describe invalid test geometry. Keep the reason for exclusion separate from the material’s measured response.

Read instrument notes, visual observations and acceptance flags together. If a spreadsheet includes both recorded and approved values, translate those categories distinctly. A technically accurate report may include failed test runs without counting them in the reported statistic.

12. Distinguish buckling from tilting

Buckling describes a deformation mode; tilting describes loss of alignment or inclination. They are not synonyms, even if both can make an edge-crush measurement unsuitable. The distinction can help the laboratory identify whether material instability or specimen positioning needs further attention.

Imagine two bilingual incident notes. One says the board bent laterally; another says the specimen tipped in the test fixture. Translating both as ‘collapse’ hides potentially different causes. Readers later reviewing photographic evidence cannot determine which phenomenon was observed.

Retain the original observation wording unless the issuing laboratory has an approved terminology glossary. Where imagery or notes conflict, escalate rather than selecting the more familiar term. A translator should not become an unappointed expert witness about the failure mechanism.

13. Separate a test result from a procurement limit

A test certificate might give a mean ECT result, while the purchase contract separately requires a minimum value. Translating the requirement as though it were an observed result can turn an unverified promise into apparent test evidence. Conversely, replacing a laboratory mean with a buyer’s limit makes the tested material appear artificially weak.

A supplier declaration says ‘minimum 5.0 kN/m’, and a lab sheet lists individual values near that threshold. The statement of supply requirement and the measurement record answer different questions. If the mean and individual observations differ, the comparison must use whatever acceptance rule the contract actually specifies.

Give ‘requirement’, ‘measured value’, ‘reported average’, ‘tolerance’ and ‘conformity conclusion’ their own consistent headings. Do not add a pass/fail stamp simply because a value looks high or low. The underlying contract and approved decision rule belong to qualified reviewers.

14. Mean, range and replicate count matter

A single result and a mean of several replicates are not interchangeable. When a laboratory gives an average, range, standard deviation or count, those statistical qualifiers help the reader assess variability. A translation that displays only the attractive headline number strips away evidence about how reliably it describes the batch.

Consider five fictional replicates of 5.1, 5.4, 5.2, 5.0 and 5.3 kN/m. Their arithmetic mean is 5.2 kN/m. That calculation is educational; it is not an ISO reporting instruction. If the original laboratory report uses another calculation convention, preserve the original rather than substituting the example.

Check that the decimal separator, replicate labels and rounding rules survive transfer between spreadsheets, languages and PDF exports. Even a correct average becomes misleading when a hidden table row, nonnumeric text cell or wrong unit is ignored.

15. Keep conditioning and environment in view

Paper-based materials respond to moisture and environmental history. Conditioning information therefore affects the comparability of reported results even when the translation assignment focuses on the strength value. It is unsafe to describe all ECT numbers as directly comparable without checking the specimen histories.

A procurement team may compare a dry warehouse sample with a board recovered from a humid distribution route. If a translator omits the conditioning note as ‘laboratory boilerplate’, the two figures appear to come from equivalent circumstances. They do not necessarily do so.

Translate the documented temperature, relative humidity, conditioning period and any deviations exactly as given. Do not insert generic standard values unless they are in the source or a referenced approved procedure. A good translation lets the expert decide which comparison is defensible.

16. Understand what uncertainty can and cannot say

Some laboratories provide measurement uncertainty while others state a result and cite an accreditation scope. Uncertainty is not a magic correction that the translator can apply to improve or downgrade a result. It describes how a measurement is reported under a defined evaluation, and its meaning depends on the laboratory’s stated convention.

A certificate has a value with expanded uncertainty and a coverage description. A marketing writer requests a shorter translated ‘guaranteed strength’ headline. Removing the uncertainty and replacing ‘reported’ with ‘guaranteed’ changes the type of claim.

Keep the value, uncertainty figure, units, coverage information and decision rule in context. If no uncertainty appears, do not invent one. If the client wants a conformity statement, request formal laboratory or quality approval before adding it.

17. Beware decimal commas and thousands marks

Some languages use a comma as a decimal separator; others use a point. In a cross-border ECT table, 5,250 may mean five point two five zero or five thousand two hundred fifty depending on context. Commas copied from numerical cells can therefore create serious ambiguity in a report read internationally.

Imagine a German-style result ‘5,2 kN/m’ inserted into a target document that normally uses decimal points. A reader may parse it correctly, yet another system may treat the cell as text. Display conventions can be adapted after verification, but the numerical value must never change.

Carry an audit column with the source string, normalized machine-readable number and target display string. Check more than one example, especially values near commercial limits. Automated localization should be tested with small values, large values and missing data.

18. Maintain the distinction between board grade and test evidence

A supplier may market a paperboard grade using an internal name that hints at strength. That name is not a measured result. The same grade can also appear in different thicknesses, flute profiles or production batches. Translators should not turn brand-like grade numbers into universal performance guarantees.

A carton drawing refers to ‘Kraft 200 / medium 150’, while a test report identifies a finished board by lot code. Those statements carry different kinds of evidence. The paper weights may help describe construction, but do not replace the ECT measurement.

Keep the supplier’s grade designation unchanged and add a translated description only when it is authorized and unambiguous. Ask whether the test actually covered that exact grade, conversion and lot before linking the reported strength to a purchase line.

19. Tell edge crush from flat crush and burst strength

Packaging reports frequently contain several physical tests. Flat crush evaluates resistance under a different loading geometry; burst strength concerns rupture under pressure; puncture or tear methods address yet other performance questions. One number cannot answer all these intents, even when marketing uses the broad phrase ‘strong carton’.

An importer sends a mixed-language table with columns ‘ECT’, ‘FCT’ and ‘burst’. A rushed translation uses one word meaning ‘compression’ for all three. This erases the distinction between test setups and makes it impossible for a customer to compare specifications correctly.

Preserve standardized test names and acronyms with a short explanation of the measured property. Match units to their respective fields. If the source uses an unfamiliar abbreviation, confirm it from the issuing lab’s legend rather than guessing based on adjacent columns.

20. Avoid a false stacking-strength guarantee

ECT is useful material evidence, but real-world carton stacking also depends on box geometry, moisture, handling, joints, openings, palletization, contents and time. A brochure may wish to convert an ECT figure into a simple sentence promising that every box will carry a specified load. That transformation is an engineering claim, not translation.

A vendor asks to turn ‘ECT measured at 6.1 kN/m’ into ‘withstanding 6,100 kilograms in a stack’. The units are incompatible and the inference is unsupported. A clear translator refuses the conversion and offers the precise material-level statement instead.

Any explanation of expected box performance should be grounded in a separate design or validation calculation accepted by the appropriate technical party. Maintain the narrow ownership of the translated result; do not build a competing packaging engineering decision system inside a language note.

21. Check test date, release date and validity period

A report can contain a sampling date, a test date, an issue date and a document revision date. These are not the same event. Translating one as another may make a client think the material was measured before it was produced or that a superseded certificate remains current.

Imagine a sample tested on 2 April, reported on 8 April and included in a supplier packet on 20 May. If the target-language heading reads ‘manufacture date: 8 April’, the chain of evidence is distorted even though every date is copied exactly.

Use unambiguous month names in multilingual documents where day/month ordering differs. Preserve time zones where stated and identify any dates that the source does not supply. The finished translation should allow a careful buyer to reconstruct the chronology.

22. Control graph labels and laboratory photographs

Charts often contain little text but carry substantial meaning. Their axes identify force, displacement or time, and photographs can show specimen geometry or deformation. Merely translating the caption while leaving the axis labels in the source language can create an incomplete technical document.

A lab plot displays ‘maximum force’ as a peak followed by a decline. A retypeset version changes the vertical scale or truncates the axis, making the difference between samples look larger. This is not an innocent aesthetic adjustment; it can alter perception of the data.

Preserve graphical values, units, legends, specimen IDs and scales. When creating localized diagrams, use the unchanged original data and perform a cross-language visual proof. Any image adjustment should be recorded as editorial, not presented as a new measurement.

23. Treat laboratory caveats as central information

Qualification statements such as ‘results apply only to submitted samples’ are not disposable legal decoration. They tell readers how far conclusions can legitimately extend. A translator who removes them may unintentionally turn limited evidence into a certification for all production output.

A supplier sends a single-board certificate with a note that sampling was performed by the customer. A sales brochure omits this detail and says an independent laboratory verified all boxes. The source does not support that larger claim.

Maintain the scope, sampling responsibility, accreditation wording and restrictions on reproduction. If marketing wants a shorter summary, craft it as a clearly bounded paraphrase that still says what was measured and what was not established.

24. Build a bilingual product specification without inventing fields

A buyer often wants one consolidated multilingual sheet with dimensions, board construction, ECT, print finish, moisture requirements and packaging instructions. Consolidation is useful, but it increases the risk that facts from different documents will be merged as though they shared one test date or one batch.

Suppose the flute profile comes from a January quotation, the strength result from a February laboratory report and the dimensions from a March drawing revision. A tidy table may conceal those different provenance chains. The translator must not choose a single date and imply every field belongs to it.

Add source-document references at field level for important technical entries. Where the documents conflict, preserve both and flag the issue. The point of the translated specification is to help procurement reach a reliable decision, not to hide inconsistencies.

25. Write an accurate customer-facing summary

Technical language may need to become readable for a purchaser unfamiliar with paper testing. The solution is not to exaggerate. A good summary explains that the ISO 3037 result concerns resistance of a corrugated board specimen to compression applied along its flute axis, using the declared method and units.

An acceptable plain-language line can say: ‘The tested board showed the reported edgewise compression resistance under laboratory conditions.’ It should not claim that a certain package will survive a drop, hold an arbitrary weight or remain unchanged in all climates.

Place the measured figure directly beside the method and source citation. Offer a link to the detailed certificate if distribution permissions allow. A shorter sentence is successful when it preserves the reader’s decision boundaries.

26. Prepare a translation sign-off trail

An industrial translation is finished only when the source evidence, target text and review decision can be reconstructed. In multilingual packaging supply chains, the document may be reused by procurement, factories, logistics providers and customer-service staff months after the initial translation.

Imagine an inspector discovers a unit mismatch after the certificate has been forwarded to three countries. Without a change log, nobody knows which version is authoritative. A controlled review record should list the source version, target version, author, subject-matter reviewer, correction date and approval.

Use a final test: can someone who reads only the translation identify the board sample, method, unit, actual result, restrictions and issuing laboratory? If yes, the translation is doing more than sounding natural; it is protecting a useful technical decision.

Three complete miniature cases for translation practice

Case A: the translated supplier quotation

A fictional supplier quotation lists a product as DW-45, states ‘double-wall board’, and gives an ECT figure of 5.6 kN/m. The annex attaches a laboratory report for lot DW-45-07 but omits the specimen date in the quotation. The correct translation separates the supplier’s quoted property from the laboratory’s documented result and keeps the lot identifier. It cannot infer that every future DW-45 order will give the same value.

A reviewer should check whether the quotation and certificate share construction, lot and edition; whether the supplier is promising a minimum or merely reporting a test; and whether the intended use requires another validation step. The translated text can be plain without making a stronger claim than the source. This is especially important when an industrial buyer is comparing suppliers through a spreadsheet.

Case B: a conversion error across language formats

A fictional laboratory spreadsheet reports 4,7 kN/m under a decimal-comma convention. During localization, a vendor’s software treats the entry as text and moves it into an adjacent column. The translated PDF shows 47 kN/m. The mistake is not linguistic in the ordinary sense, yet the result is false because the language workflow lost a separator. A good review compares source strings, numeric values and final displayed units.

The error can be prevented with machine-readable numerical cells, locale-aware formatting and one-to-one verification of the certificate table. Ask the reviewer to recalculate a sample line and confirm that the chart annotations match the corresponding numeric columns. The production document should be rejected until the value is corrected, even if the prose reads fluently.

Case C: an unsupported stacking promise

A fictional online listing says the board has an edge crush value of 6.0 kN/m under the named laboratory test. A marketer wants to translate the statement into a claim that a loaded carton can be stacked ten high for a month. The source provides no box dimensions, contents, load history, temperature or distribution simulation. The additional claim is not a translation; it is an unvalidated change to the product promise.

The appropriate response is to translate the measured edge crush statement exactly, and treat any further packaging-performance sentence as a separate engineering question. Readers benefit from knowing where the test ends. That boundary is a mark of clarity rather than a lack of confidence.

A four-week practical learning plan

Week 1 — decode the object and method

Collect three anonymized packaging records. Label every reference to board, sheet, specimen and finished carton. Identify the test method and edition in each, and explain aloud why an edge-crush result is not a complete box-compression declaration. Write one short bilingual glossary focused on the specimen and loading direction.

Week 2 — master numbers and document identity

Extract ten values with units from certificates and quotations. For each, distinguish raw force, normalized resistance, average, uncertainty, target and threshold. Build a source-to-target value table. Reformat the numbers in two locale conventions and test whether an independent reader obtains exactly the same quantities.

Week 3 — practise difficult commercial cases

Write two realistic supplier emails, one involving a missing sample code and another involving inconsistent editions. Draft neutral clarification questions rather than inserting guesses. Translate a sample customer-facing summary that communicates what was tested while retaining the limitation on complete-box performance.

Week 4 — complete a controlled review

Simulate the handover of a four-page report containing a figure, table, certificate note and purchase specification. Ask one bilingual reader and one technical reader to review independently. Fix discrepancies, record decisions, prepare the revision log and verify that every key measurement links to original evidence.

Frequently asked questions

What does ECT mean in this guide?

ECT is commonly used for an edge-crush measurement of corrugated board. Its exact test and reporting context must be read from the cited laboratory document rather than assumed from a marketing acronym.

Is the ISO 3037:2022 result a box stacking rating?

No. It concerns a test on corrugated fibreboard specimens. Predictions or declarations concerning finished boxes require their own appropriate evidence and design assumptions.

Can I translate kN/m as kilograms?

Not as a direct substitution. kN/m expresses force per length. Kilograms describe mass, and no direct product-load equivalence follows from the ECT value alone.

Does the flute direction matter?

Yes. The public ISO scope identifies compression in the direction of the flute axis. Preserve specimen orientation and related diagram information.

Can a report omit invalid runs?

Translate the issued record faithfully. If the original distinguishes excluded tests, retain that distinction; do not silently invent or suppress data.

Should I update an older ISO citation to 2022?

No. Preserve the cited edition in the source and let the issuing party or appropriate expert determine whether revised testing is required.

How should I translate double-wall board?

Use a target-language construction term that identifies two corrugated mediums, subject to the terminology convention of the actual document. Do not reduce it to generic ‘double thickness’.

What if the quoted ECT differs from a lab report?

Keep requirement, declaration and measured result separate. Request a reconciliation based on material code, lot, sampling details and the governing contract.

How can a translator check a decimal comma?

Confirm the numerical value with the source table and apply consistent locale conventions. Verify the final exported PDF, not just the editable document.

Who approves the final engineering claim?

The authorized laboratory, manufacturer or qualified technical owner should approve performance or conformity claims. Translation review checks faithful communication, not independent certification.