���ϴ��¼�

The degree to which a translation correctly communicates the meaning of the source content without adding, omitting, or changing information. Accuracy is considered alongside fluency, terminology, style, and fitness for purpose.

The modification of translated content to make it appropriate for a particular audience, culture, channel, or use case. Adaptation may involve changing examples, measurements, visuals, references, tone, or messaging rather than translating the source literally.

The process of matching corresponding source and translated sentences or segments. Alignment can convert previously translated documents into a translation memory so approved content can be reused.

The use of artificial intelligence to produce or support translations. It can include neural machine translation, large language models, automated terminology application, quality estimation, content classification, and intelligent workflow routing.

An application programming interface, or API, allows software systems to exchange information and trigger actions automatically. A translation API can submit content, retrieve completed translations, and connect translation workflows with content or product systems.

Automated quality estimation, often abbreviated as AQE or QE, uses a model to predict translation quality without comparing the output with a professionally created reference. It can help identify content likely to require human review.

The translation of completed target-language content back into the original source language, usually by an independent linguist. It can help identify changes in meaning in clinical research, regulated communications, surveys, and patient-facing materials.

Bidirectional, or BiDi, text contains both left-to-right and right-to-left writing. Arabic or Hebrew content may include English names, numbers, URLs, or codes that require careful directional handling.

A file that displays or stores source and target content together. Bilingual formats allow translators and reviewers to work with aligned segments while preserving tags, formatting, and metadata.

The consistent personality, tone, vocabulary, and communication style used by an organization. Multilingual brand voice guidance helps preserve a recognizable identity across markets.

A computer-assisted translation tool is software that helps professional linguists translate and review content. It provides segmented content, translation memories, term bases, quality checks, comments, and reference materials.

A translation accompanied by a signed statement confirming that it is complete and accurate to the best of the translator’s or language service provider’s knowledge. Requirements vary by country, institution, and document type.

The system used to represent letters, symbols, punctuation, and other characters digitally. Incorrect encoding can cause multilingual text to display as boxes, question marks, or corrupted characters.

Software used to create, organize, publish, and update digital content. Connecting a CMS to a translation platform can reduce manual file handling and support continuous multilingual publishing.

The surrounding information needed to understand what a word, sentence, or string means and how it will be used. Screenshots, previous sentences, character limits, audience information, and interface location can all provide translation context.

A process that integrates translation into ongoing content or software development. New or changed content is identified, translated, reviewed, tested, and delivered regularly rather than through occasional large releases.

A restricted form of language that uses approved vocabulary, grammar, and sentence structures. It can reduce ambiguity, improve readability, and support stronger AI and machine translation performance.

A structured collection of written or spoken language used for linguistic analysis, model development, terminology research, or translation reference. A parallel corpus contains corresponding content in multiple languages.

The formatting and layout work required to produce a polished translated document. Multilingual DTP may address line breaks, tables, typography, page flow, images, text expansion, and right-to-left content.

A regional or social form of a language with distinctive vocabulary, grammar, or pronunciation. Translation requirements should identify the intended regional audience when dialect differences affect meaning or tone.

A word, product name, code, trademark, or expression that must remain unchanged in target content. These terms should be recorded in the project glossary for consistent use by translators, reviewers, and AI systems.

The review of a translation against the source content to improve accuracy, terminology, completeness, style, and readability. Editing is often completed by a second linguist after the initial translation.

A classification indicating the effect a translation issue may have on meaning, usability, compliance, safety, or brand perception. Programs commonly classify errors as minor, major, or critical.

A translation memory result where the source segment is identical to stored content. Exact matches may still require contextual review because the same sentence can serve a different purpose in a new location.

A word that looks or sounds similar in two languages but has a different meaning. False friends can create convincing errors because the incorrect choice initially appears familiar.

Testing that verifies whether a localized website, application, or digital product works correctly. It may cover links, forms, buttons, navigation, data entry, language switching, and text direction.

The degree to which translated content reads naturally and correctly in the target language. Fluency reflects grammar, sentence structure, word choice, punctuation, and style.

A translation memory result that is similar, but not identical, to a new source segment. A linguist reviews and modifies the proposed translation to account for differences in content and context.

The connected processes used to create, govern, translate, review, publish, and maintain content across markets and languages. A mature model connects people, technology, language assets, quality controls, and business systems.

The overall process of preparing an organization, product, or content strategy to operate across international markets. It can include internationalization, localization, global brand management, market readiness, and regulatory planning.

A structured collection of important terms and their meanings. In translation, a glossary usually includes approved target-language equivalents, definitions, context, usage notes, and prohibited variants.

The framework of policies, roles, approval rules, quality standards, security controls, and reporting practices used to manage multilingual content.

A workflow that combines automated translation with professional human involvement. Linguists may verify meaning, correct terminology, improve fluency, validate high-risk content, or approve final delivery.

Translation performed by a professional linguist rather than generated automatically. Human translators interpret context, intent, tone, cultural references, and subject-specific requirements.

A workflow combining methods such as AI translation, translation memory, professional human translation, post-editing, subject-matter review, and automated quality checks.

Evaluation by a reviewer who works in or closely understands the target market. In-country reviewers may confirm local terminology, regulatory wording, brand preferences, product knowledge, or market appropriateness.

The design and preparation of software, content, or products so they can support multiple languages and locales without extensive redevelopment. It may include externalizing text, supporting Unicode, and enabling locale-specific formats.

An international standard for translation services that defines requirements for translator qualifications, project management, translation, revision, review, quality assurance, and related professional processes.

An international quality management standard focused on documented processes, customer requirements, risk-based thinking, performance measurement, corrective action, and continual improvement.

The communication of spoken or signed language from one language into another. Common forms include simultaneous, consecutive, over-the-phone, video remote, and on-site interpreting.

JavaScript Object Notation is a structured data format widely used in software, websites, applications, and APIs. Localizable JSON files may contain interface text alongside keys, variables, and syntax that must remain intact.

The translation and adaptation of help articles, product documentation, troubleshooting instructions, and customer support content. Effective workflows keep localized versions synchronized with changing source content.

Reusable linguistic information that supports translation quality and consistency. Translation memories, term bases, style guides, corpora, approved reference translations, and do-not-translate lists are common language assets.

A standardized identifier for a language, such as en for English or fr for French. Extended codes can identify regional variations, such as en-US or pt-BR.

The source and target languages used for a translation, such as English-to-German or Japanese-to-English.

A company that delivers translation, localization, interpreting, terminology, testing, and related language services. Enterprise LSPs may also provide technology integration, governance, security, and program management.

An AI model trained on extensive language data to understand and generate text. LLMs can support translation, rewriting, terminology analysis, content classification, and quality evaluation.

The structured evaluation of translated content for accuracy, fluency, terminology, grammar, style, consistency, locale conventions, and project requirements.

A controlled process used to confirm that translated content communicates the intended concepts appropriately in the target language and culture. It may include forward translation, reconciliation, back translation, cognitive debriefing, and final review.

A defined combination of language, geography, and cultural conventions. It can determine wording, date formats, currencies, numbers, addresses, units of measure, and interface behavior.

The adaptation of content, software, or a product for a specific market. It includes translation and may also address cultural relevance, formatting, imagery, legal requirements, and interface behavior.

Testing that verifies whether a localized product works and appears correctly in its intended market. It may evaluate language, layout, truncated text, input fields, locale formats, character display, and cultural appropriateness.

Software-generated translation from one language into another. Current systems commonly use neural networks, large language models, or other AI techniques, and quality varies by model, language, domain, context, and content type.

The system or model that produces automated translations. Engines may be general-purpose, domain-adapted, customized with client data, or optimized for particular languages and content types.

The human review and correction of machine-generated translations. Post-editors evaluate meaning, terminology, fluency, formatting, and adherence to project requirements.

The adaptation of promotional content for audiences in different languages and markets. It often requires attention to brand voice, persuasion, cultural relevance, search behavior, slogans, and channel requirements.

Translation of clinical, pharmaceutical, healthcare, medical device, patient-facing, and scientific content. It requires qualified linguists, subject knowledge, terminology control, and appropriate review based on risk and intended use.

A framework for categorizing and evaluating translation errors across dimensions such as accuracy, terminology, fluency, style, locale conventions, and formatting.

The adaptation of digital content so it can be discovered through search in different languages and markets. It may include local keyword research, translated metadata, localized URLs, internal linking, hreflang, and regional search intent.

A translator, editor, or reviewer with native-level command of the target language and deep familiarity with its cultural and regional conventions. Professional work also requires subject knowledge, writing skill, research ability, and quality discipline.

Machine translation that uses neural networks to generate output based on patterns learned from large amounts of bilingual and multilingual data. NMT can produce fluent language but may still mishandle context, terminology, names, or numbers.

Content intentionally kept in the source language, such as product names, trademarks, codes, model numbers, legal names, software commands, or standardized terminology.

Technology that converts text contained in scanned documents or images into editable, machine-readable content. OCR output should be checked carefully before translation, especially for complex layouts or regulated information.

A symbol, code, or variable that will be replaced with dynamic content when software runs. Placeholders must usually remain unchanged and in the correct position during translation.

The human correction of automatically generated translation. It may focus on essential meaning and usability or require a fully polished result comparable to professional human translation.

The final review of translated content for typographical errors, punctuation, grammar, formatting, and presentation. It is usually performed after translation and editing.

The replacement of source text with simulated localized text to test whether software is technically ready for translation. It can reveal hard-coded text, character corruption, layout restrictions, and right-to-left issues.

An instruction or set of contextual information provided to an AI model. In translation, a prompt may specify language, audience, terminology, style, domain, formatting, or quality expectations.

The system of processes used to prevent, identify, and correct problems before translated content is delivered or published. Translation QA can combine qualified linguists, independent review, automated checks, terminology controls, testing, and corrective action.

A measurable result based on the number, type, severity, or weighted impact of review findings. Scores are most useful when teams apply consistent evaluation criteria.

The minimum acceptable standard for a translation or workflow. Different thresholds may be established for internal reference content, customer communications, marketing, legal documents, or safety-critical materials.

A form of a language used in a particular country or region. Variants can differ in spelling, vocabulary, grammar, tone, date formats, and cultural expectations.

A source segment that appears more than once within a project or set of files. Translation technology can identify repetitions so approved translations are reused consistently.

A file that stores translatable software content separately from application code. Common formats include JSON, XML, YAML, properties files, and platform-specific resource formats.

The direction of translated content to the appropriate translator, editor, subject-matter expert, in-country reviewer, or approver based on defined rules.

A writing direction used by languages such as Arabic and Hebrew. Supporting RTL content can require mirrored layouts, adjusted navigation, correct alignment, and bidirectional-text testing.

A unit of source content processed within a translation environment. It may be a sentence, heading, list item, interface string, or another defined portion of text.

A definition of measurable service commitments between a provider and client. Translation SLAs may address turnaround, response time, availability, quality, issue resolution, security, reporting, and escalation.

The original material submitted for translation or localization. Clear, complete, and well-structured source content generally produces better translations and reduces questions and rework.

The original language from which content is translated. A project may have one source language and multiple target languages.

A sequence of characters used as content within software, websites, or digital products. Strings may contain visible text, placeholders, markup, variables, and character restrictions.

A document defining writing and formatting preferences for multilingual content, including tone, voice, punctuation, capitalization, units, dates, product naming, and locale conventions.

A person with specialized knowledge of a product, discipline, industry, or business function. SMEs may clarify source content, approve terminology, validate technical meaning, or review high-risk translations.

The presentation of translated or transcribed dialogue as timed on-screen text. Effective subtitles must communicate meaning within limits for reading speed, line length, timing, synchronization, and screen space.

The language into which content is translated. The intended locale and audience should also be identified because one language may have multiple regional variants.

A structured, searchable database of approved terminology. Entries may include definitions, translations, context, subject fields, usage notes, grammatical information, status, forbidden terms, and examples.

The identification of important or frequently occurring terms within source content. Candidate terms are reviewed, defined, translated, and approved before being added to a glossary or term base.

The systematic process of identifying, defining, translating, approving, distributing, and maintaining important terms across languages and workflows.

The communication of written meaning from a source language into a target language. Professional translation considers accuracy, context, purpose, terminology, tone, audience, cultural expectations, and delivery format.

An interface that connects translation capabilities with websites, applications, content systems, or internal business tools. It can automate content submission, project creation, status updates, delivery, and repetitive workflow steps.

A collection of approved source terms and their target-language equivalents. It may also contain definitions, context, usage instructions, prohibited translations, grammatical information, and approval status.

A platform used to coordinate translation projects and multilingual content workflows. A TMS may manage files, users, vendors, translation memories, term bases, automation, reviews, quality checks, reporting, integrations, and delivery.

A database that stores previously translated source and target segments so identical or similar content can be reused. Translation memory can improve consistency, shorten turnaround, and reduce repetitive work.

The combination of people, processes, technologies, and controls used to verify that translated content meets defined requirements.

A stored pair consisting of a source segment and its corresponding translation. Translation units form the core of a translation memory and may include project, date, language, creator, and approval metadata.

The creative adaptation of messaging for a new language and culture while preserving the original communication objective and emotional effect. It is commonly used for campaigns, slogans, and brand messaging.

The conversion of spoken audio or video into written text. The transcript may remain in the original language or become the source for translation, subtitling, captions, voice-over, or analysis.

The representation of words from one writing system using the characters of another. It focuses on how words are written or pronounced rather than translating their meaning.

An international character-encoding standard designed to represent text from nearly all writing systems. UTF-8 is one of the most widely used Unicode encoding formats.

The adaptation of visible and interactive software elements for a target market, including buttons, menus, navigation, messages, forms, notifications, help text, locale formatting, and right-to-left behavior.

Confirmation that translated content is appropriate for its intended purpose and meets defined requirements. Validation may include linguistic review, subject-matter approval, regulatory confirmation, user testing, or source comparison.

The process of selecting, qualifying, assigning, monitoring, and developing external translators, reviewers, interpreters, and language service providers.

The adaptation of a website for users in different languages and markets. It includes content, navigation, metadata, forms, images, user journeys, dynamic content, multilingual SEO, and technical delivery.

A method used to estimate translation volume. Counts can vary by file type, language, extraction method, and whether repetitions or previously translated content are included.

The use of rules and system integrations to reduce manual translation tasks. Automation can identify changed content, create projects, assign resources, trigger reviews, perform checks, and return completed translations.

XML Localization Interchange File Format is an industry-standard bilingual format designed to exchange translatable content while preserving source text, translations, tags, status information, and metadata.

Extensible Markup Language is a structured format used to store and exchange data. XML files may contain translatable content alongside elements, attributes, codes, and technical instructions that must be protected.

Translation performed by an AI model for a language pair or task without task-specific training or adaptation. Results may vary significantly for lower-resource languages, specialized terminology, and high-risk content.