Real-World Terminology Management - TermNet - International

1

Real-World Terminology Management

Sue Ellen WrightKent State University

Terminology Summer School

©Sue Ellen Wright 2 of 44

Acronyms

TMM=Terminology ManagementTWF=Terminology WorkflowKOS=Knowledge Organization SystemSKOS=Simple KOSOWL=Web Ontology Language

i18n=InternationalizationQA=Quality AssuranceSL=Source LanguageTL=Target LanguageISO=International Organization for StandardizationROI = Return on Investment

2


Terminology Management Spectrum

Continuum of practice in a range defined by multiple factors:

Enterprise type and missionCriticality of terminology for core businessCriticality of quality and branding concernsRecognition of tangible and intangible ROITerminology user groupsBuy-in by stakeholders (design, engineering, marketing, as well as client & in-country partners, but above all, top-level management)


Enterprise Type

Nature of enterpriseGovernmentIndustryResearch institutesLocalization/translation bureaus

Nature of the clientText typeNegotiated client/vendor agreements

Web content management providersFreelance & in-house technical writers, translatorsLibrarians & knowledge organization environments

3


TMM Mission & Strategic Position

Support for national language policySupport for global enterprise activitySupport for technical writing & translation

Ongoing, stable subject fieldsFrequent one-off, unrelated, unanticipated subject fields

TMM approachesPrescriptive vs. descriptiveAd hoc vs. systematic


TWF Task Issues

Specific task typesLanguage planning StandardizationDocument productionControlled language document productionActivity in localization & multilingual documentation environmentsSupport for machine vs. human-oriented translationContent management in dynamically changing Web environmentsTerminology management for enterprise solutions (whatever that is)

4


Knowledge-oriented TWF

Enterprise-related information and knowledge management

Terminology as a function of taxonomy, ontology, and information retrieval (knowledge organization systems: KOS)KOS interaction with:

Monolingual technical writing & product developmentTranslation and localizationMultilingual technical writingE-business solutionsInventory control and logisticsGeneral information retrieval and processingStandards issues (SKOS and OWL)


Cost Effectiveness:

Terminology management activity costs time and money. Is it economically feasible:

For individual translators?For groups working together?For bureaus and translation services in government and industry?

5


Criticality of Terminology

Potential for market lossesPotential for communicative losses Risk of product failure, human injuryAdverse effects on branding efforts (marketing issues)Relative significance of terminology

To the processTo the productExample: Terminology is more criticalif you are selling software than if youare selling wheat.


ROI: Variable vs. Fixed Costs

TMM tasks hidden in general overheadCost of not agreeing to uniform terminology hidden in overhead costsCosts of correcting mistakes or recouping damage to branding difficult to documentCompounded costs due to the persistence of defective communications

6


Fixed Costs

The costs of doing systematic terminology work, on the other hand, are up front costs that can be calculated usingtraditional means.


Calculating Return on Investment

Greater general applicability of specific terminological units (greater frequency) = greater the return on terminology management costs. Greater the quality, safety, or competition-related criticality = greater the return.The greater the degree of integration between straight CAT, TM and MT applications, the greater the payback in leverageable data.

7


Terminology Management Spectrum

These factors result in varying:Integrated and non-integrated workflow modelsComplexity with respect to the data model Investment of human resources in the form of

Term extractionContext documentationConcept definition (ideally by SL & TL subject experts)Terminology product delivery

– On-screen lookup– Web-based resources– Hard-copy output


Complexity of Data Model

A continuum of practice in a range defined by:Creation of non-systematic tab-delimited, text and job-oriented glossaries and spreadsheets

SL term + TL term + POS, Gender, and/or a note

Theoretically rich complex entry input modelsSubject field classificationClient designationConcept-oriented definitionSL term + TL term + POS + Term Type + Register + Context, etc.Documentation through source citations, graphics, etc.

8


Excel Spreadsheet Glossary


Highly Granular Input Model

9



10


What is workflow?

Workflow is the operational aspect of a work procedure:

how tasks are structuredwhat their relative order ishow they are synchronizedwho performs them (& where)how information flows to support the tasks and how tasks are being tracked

http://en.wikipedia.org/wiki/Workflow


Macrostructures & Task Structures

Terminology workflow (TWF) as a task-set in overall enterprise workflowTerminology management (TMM) mission within the enterpriseWorkflow parametersTask issuesInput issues

11


Evolving Workflow Models

TWF as non-detailed item in overall workflowTWF plotted as task set during project planningTWF plotted independently for individual projectsTWF plotted for ongoing activities involving ongoing global TMMProject-oriented workflow vs. ongoing global enterprise-related terminology management


Critical Workflow Parameters

Dimensions:Tasks (activities)

Time–How much time does it take?–How much time do you have?

ThroughputResources

HumanTools“Raw Materials”

12


Critical Workflow Parameters

Dimensions:Cost accounting as a function of workflow management

Difficult to calculate because of individual task differencesBalance between intensity & granularity of TMM vs. the investment costs incurred for “doing” TMMBalance between investment and incidence of reuseBalance between investment and criticality of terminological accuracy


Project Inputs

Raw source textFrequently flawed (particularly in localization environments)Subject to ambiguity: polysemy, synonymyTerm extraction from the source text

Human or automatic term recognitionSynonym identification and disambiguationCo-occurrence issues with advanced automatic indexing software

Determining target language equivalentsVerification issues as a function of workflow

13


Ordering and Synchronization

Traditional TMM workflow positionRationalized workflow positionGlobal TMM solutionsSystematic TMMInformation feedback loop


Traditional Position of Terminology Management (TMM) in Global Project Workflow

Ad hoc TMMReactive project-specific TMMNo influence on document production, i18n

End-ItemInspection

TerminologyManagement

Workflow diagram ©Keiran Dunne 2005

14



Rationalized Project-Oriented TMM

TMM as a function of QA (Quality Assurance) managementTMM and QA upstreamed to planning stageProactive TMM

Workflow diagram ©Keiran Dunne 2005


Global Enterprise Terminology Management Solutions

CorpusManagement

Corpus Generation & Maintenance

TerminologyExtraction

Term Proposal


Input

Modification

Check

Verification

Approval

Origination

Publication

Suggestion

Etc.

Use

15


Terminology Management Transactions (ISO 12620)

Proposal of a term / extraction of a termOrigination: creation of a term entryInput: data inputModification: update and changeCheck by terminologist Verification by subject field expertsApprovalWithdrawal“Publication”Use “in the field”Modification suggestions(Standardization and language planning)


Task Definitions

Origination: entry creationInput: population of data fieldsModification: update & correctionCheck: terminology checkVerification: subject field check (in-country reviewer?)Approval: sign-off by chief terminologistPublication: importation into master database, upload to Web, Internet, intranet, LAN, hardcopy, etc.

16


Systematic Term Management for Collections of Controlled Size

CorpusManagement

Corpus Selection

Concept Identification

Concept Selection


DefinitionCreation

Entry Creation(see previous)

TermAssignment

Term StatusAssignment

ConceptSystem

TermCreation

Language Planning

Prescriptive Terminology


Dynamic Systematic Management for Large Collections

CorpusManagement

Corpus Collectionand/or Selection

Term/Concept Identification

Concept Field Analysis


DefinitionCreation

Entry Creation(see previous)

TermAssignment

Term StatusAssignment

ConceptPositioning

TermCreation

DynamicKOS generation

KOS=Knowledge Organization System

OntologyLink

Ontology

(optional)

17


Information Feedback Loop

ProjectPost Mortem

Reviewer Corrections

SuggestionReceipt

TerminologistUpdate

Procedure

Master TDBPublished

TDB

User Suggestions

Research

Verification

Approval

UpdateData Entry


Who Does What

Project developers (engineers, designers, other subject experts)

Terminological inconsistenciesUncontrolled coinage of neologisms

Terminology project group representativeTechnical writers

Frequently unaware of multilingual issuesLanguage specific problems

TranslatorsKnowledge engineersTerminologists

18


Who Does what

Master terminologistAuthority to change records in master fileAuthority for final approval

Language-specific terminologistSource and target language research

Data input specialistSubject-field specialist

Source languageTarget language

Conflation and expansion of tasks according to need


Where do they do it?

Workflow solutions and application design dependent on work group configurations:

Individual terminologist in one locationSmall work groupSingle local work groupLarger distributed networkMultilingual globally distributed network

19


Publication, Dissemination and Access

Local master terminology databaseWeb-based terminology resource

Web-serverEnterprise-based IntranetLAN server

Exportation to other TDB environmentsHardcopy

Special project sub-setting


Special Outputs

Project-specific term entry subsetting and exportation (electronic)Selective subset delivery, especially to freelance developers, tech writers, and translatorsOutput of SL TL glosses in hard copy

For subject area specialists in multilingual environmentsFor interpreters

20


Delivery to Technical Writers and Translators

Translators: search for equivalent terms based on source language terms and conceptsTechnical writers: search for source language terms based on developers’ concepts

Frequently flawedFrequently newly coined terms and concepts

Delivery to and buy-in from technical writers is very difficult to achieve.


Solutions for Technical Writers and Program Developers

Ontology-enabled terminology deliveryStyle sheetsAutomatic term use checkersIssues:

Resistance to synonymy controlCoinage of translation-unfriendly termsCoinage of “cute” unmotivated terms

21


Translation-oriented Solutions

Terminology database integrated with translation memory and text production softwareConcordance features for access to terms in contextMultiple layers of information

Data category subsetting for rapid information retrievalOptional view of full terminological entries


Tracking Issues

Workflow-related database design issuesTransaction-related data categories

Terminology transactionResponsibilityDateAuthorizationsTask-related check-offsOutput and sort-related categoriesImport / export trackingVersion tracking

22


Computer-mediated Workflow Management Systems

SDL WorkFlow Service™Trados Global Content Management™Star Translation Workflow Server™Lionstream Workflow™LTC Organiser Multilingual Management and Workflow Control Software System™

Non-exhaustive list!


For More Information

Sue Ellen WrightInstitute for Applied LinguisticsKent State University109 Satterfield HallKent, Ohio 44242, USA

[email protected]

Real-World Terminology Management - TermNet - International

Documents

Transcript of Real-World Terminology Management - TermNet - International