Prof Klaus: Terminology Management

Terminology Management
in Technical Communication

tcworld India
Bangalore, 11 – 12 March 2011

Klaus-Dirk Schmitz
Institute for Information Management
Faculty 03
University of Applied Sciences Cologne
klaus.schmitz@fh-koeln.de

Overview

 What is terminology?
 Basic principles of terminology management
 Terminology management in companies
 Tools for managing terminology
 Terminology extraction tools
 Terminology databases
 Terminology control tools

K.-D. Schmitz, IIM, FH Köln

What is terminology ?
Note: If you omit the password, MultiTerm
prompts you for a password when loading,
assuming the database is password-protected.
If you log on as the system administrator, you
are normally asked whether you want exclusive
access to the database. This is not the case when
opening a database using parameters; in this
case, it is assumed that you do want exclusive
access. Only when exclusive access is not
available, MultiTerm does assume that you still
want to take part in normal multi-user operation.


What is terminology ?

database terminology
exclusive access =
loading vocabulary of
log on a subject field
MultiTerm * =
multi-user operation Gesamtheit der Begriffe
open a database und Benennungen in
parameter einem Fachgebiet
(DIN 2342)
password
=
password-protected
set of designations belonging
prompt to one special language
system administrator (ISO 1087-1)

Terminological Triangle

concept

„mouse“
term object

Object

 Any part of the perceivable or
conceivable world

 Objects may be material (e.g. mouse)
or immaterial (e.g. magnetism)


Concept

 Unit of thinking
made up of characteristics
that are derived by categorizing objects
having a number of identical properties (DIN)

 Unit of knowledge created by a unique
combination of characteristics (ISO)
 Concepts are not bound to particular languages.
They are, however, influenced by social or cultural
background

“mouse”
Term

 Designation of a defined concept
in a special language
by a linguistic expression

 Designation: Any representation of a concept

 A term may consist of one or more words A term may
consist of one or more words
Single word term: mouse, printer, laser
Multi-word term: laser printer, printer with single-sheet feed


Synonymy

 Communication can be disturbed!

“return key?”

“enter key”


Synonymy
Synonymy exists
if two or more
terms in a given
language
represent the
same concept.


Synonymy


Homonymy / Polysemy
 Communication can fail!

“mouse”


Homonymy / Polysemy
Polysemy: etymological affinity,
the same word.
Homonymy exists if one term or
several terms that have the
same external form refer to
several concepts.
True homonymy: different
words with the same form, no
etymological affinity.
Differentiation between
polysemy and homonymy is
irrelevant for terminology work.

Coining new terms


Term-related issues

 Coining of terms: word formation + term buildings mechanisms

 Composition: cyberspace, translation memory system

 Derivation: preface, management

 Conversation: the chair  to chair, green (adj)  the green

 Terminologization: mouse (IT)  mouse (bio, general),
virus (IT)  virus (med)

 Loan word: festschrift, zeitgeist from DE, rickshaw from JP

 Abbreviation: CEO, AIDS, scuba, Interpol

 New creation: blurb, quark (very rare)

Selecting “good” terms

• USB flash drive
• flash drive
• USB stick
• USB memory key
• memory stick
• keydrive
• pendrive
• thumbdrive
• jumpdrive
• etc. K.-D. Schmitz, IIM, FH Köln
© Prof. Dr. Petra Drewer & Prof. Dr. Klaus‐Dirk Schmitz

Term-related issues

 Criteria for the selection and creation/coining of
terms:
 Transparency/motivation
 Consistency
 Appropriateness
 Linguistic economy
 Derivability
 Linguistic correctness
 Preference for native language

Transparency

 The concept designated by the term can be
inferred without a definition
(e.g. pipe wrench or adjustable wrench vs. monkey wrench)

 The meaning of the term is visible by:
 morphological motivation
page setup, error message
data network identification code*
network printing device setup*
 semantic motivation
worm, virus, infected file, vulnerability, firewall*


Consistency

 Terminology must be defined accurately and
used consistently at least within:
 one document
 one product

 one company or organization

 Only one term for each concept
(avoid synonyms !)
 Only one concept for each term
(avoid homonyms !)
 But also consistent term coining (see Wikipedia: wrench)


Appropriateness

 Appropriateness means that terms:
 have to be familiar to the user (localization!)
 don’t cause confusion or insecurity
 have no negative connotations (neutral, politically correct)

 express installation (only components needed)
network installation (all components)
 system error, severe error, fatal error, user error etc.
 master/slave, web designers’ bible, knowledge nugget etc.
 nuclear energy vs. atomic energy


Appropriateness


Other features
 Linguistic economy
 Ultrakurzwellenüberreichweitenfernsehrichtfunkverbindung
 Derivability
 medicinal plant vs. herb  herbal, herbalist, ...
 Bedeutungslehre  Semantik
 Linguistic correctness
 aktualisieren vs. updaten, geupdated, upgedatet, ...
 OpenSource, Cafe ToGo, fünfköpfiger Familienvater, …
 Preference for native language
 Startseite vs. Homepage, Multifunktionsleiste vs. Ribbon


Terminology in companies
 Terminology is an important carrier of knowledge for
domain-specific information in companies
 Within a company, but also between company and
customers as well as between company and suppliers
 Many departments and sectors of a company create,
disseminate, retrieve and use terminology

 BUT: Terminology management is discussed controversially:
 Costs + effort

 Quality and efficiency

 Therefore: Costs and benefits of terminology management


tekom survey
 Successful terminology
management in companies
Practical tips and guidelines:
Basic principles, implementation,
cost-benefit analysis, system
overview

published in German 4/2010,
about 300 pages, CD with data

also available in English


tekom survey
 Online questionnaire end of 2009
 about 1,000 sent out, mostly to tekom members
 High response rate of 940 questionnaires (77% tekom)
 34% managerial staff and CEOs
64% employees
 67% industrial enterprises
15% software companies
13% service providers (TD / translation / localization)
 (And: questionnaire for tools providers, questionnaire for
benchmarking companies (25), 2 benchmarking workshops)


tekom survey: the situation
At an average, a company

 has to manage 11.81 different information products

 is creating 5.87 different technical documentations

 has to deal with the fact, that 5.04 different sections/
departments are involved in the process of creating
(new) terms

 is translating information products into 10.1 different
languages

Product planning and R&D & Product Management  Product specifications
development R&D  Procedural instructions
 Process documentation for product development
 CAD graphics, 3D-models pertaining to the product
 Laboratory manuals
R&D & software development S oftware GUIs / user menus
R&D & TD marketing  Images pertaining to the product
R&D & TD  Data sheets
 Parts lists
R&D & Marketing  Packaging labels / product labels
QM  Quality documentation

Product marketing R&D & Product management  Information pertaining to product release changes
Marketing and product management  Product information sheets for pre-sales
Marketing  Marketing material
 Customer presentations
Sales and Marketing &TD  Product catalogues
 Price catalogues
 Contract documents
Corporate communications  Press releases
Product usage Technical Communication (TD)  Montage and installation instructions
 Commissioning manuals
 Online help
Training & TD  Training documents
Marketing and TD  Multimedia/simulation programs
TD & R&D & software  dev. User interfaces(GUIs) / software descriptions
R&D  Control elementsand labels
Product maintenance TD  Maintenance/S ervice manuals
TD & Service R epair manuals
S pare parts catalogues K.-D. Schmitz, IIM, FH Köln

Who is creating terminology?
Multiple answers,
Technical documentation 79.7%
average 5.04
Research / (software) development / 79.7%
engineering different
Marketing 63.5% departments/
Product management/ Portfolio 61.3% sections
management
Translation / Localization 40.4%
Distribution / Sales 39.3%
Customer service / After sales 30.7%
Training 28.4%
Management board 26.3%
Corporate communications / Public 24.4%
relation
Quality assurance / Quality management 16.3%
Purchase / Procurement 12.3%
Montage / Assembly planning / 12.3%
Production
Servicing / Maintenance 8.2%
IT service 6.7%

Terminology problems
 84.2 % report, that always or very frequently various
departments/sections use different terms for the same
concept

 70.7 % report, that always or very frequently differing
terms for the same concept are used in various
documents

 47.1 % of the staff always or very frequently have
problems in understanding technical terms on the spot

 51.1 % of the staff always or very frequently have to
ask for or retrieve the correct term for a given concept

Consequences: Terminology problems


Opinions about terminology
96,9%
96,0% 96,0%

87,7%

sehen die Zeitersparnis in halten die
Making work
schätzen die
Improving Better
gehen von einer eher großen
Saving time Arbeitserleichterung für eher Qualitätsverbesserung von bis sehr großen Erleichterung
der Kommunikation und
understanding
Arbeit als eher groß bis sehr easier
groß bis sehr groß quality
Dokumenten und in der der Verständlichkeit für den
for the customer
groß an Kommunikation als eher Kunden aus
groß bis sehr groß K.-D. Schmitz, IIM, FH Köln

Terminology documentation

Documentation of terminology

Definitions 84.3%
Subject field information 78.2%
Status (preferred, admitted, deprecated, do not use etc.) 72.3%
Grammatical information (Gender, POS, Number etc.) 51.4%
Project, product, customer, department information 44.9%
Illustrations 34.8%

 And: Context, Example, Synonym, No-Term, Source, Explanation,
References, Position numbers, short forms


Model for a cost-benefit analysis

1. Analysis of the problem
2. Analysis of the impact (very often, very expensive, bad quality)
3. Framework conditions for the value of benefit (many
documents, high volumes, many languages, using TMS, using CMS)
4. Definition of goals (management triangle: costs, time, quality)
5. Analysis of benefit (consistent terminology, less changes, higher
TM match rate, less queries by translators)
6. With/without comparison
7. Analysis of costs (initial costs: system, implementation, training;
running costs: licenses, personnel, working hours)
8. Cost-benefit analysis
9. Success factors and risks (early, involve all, workflow)

Key performance indicators: benefits
For a specific application scenario:
 Costs for changes of terms in the source language
 Costs for queries from translators
 Costs for terminology related translation corrections
 Translation costs
e.g. for changes of terms:
 duration in number of hours of work for making changes
 wages for the hours of work
 number of changes made per document
 number of newly created documents per year

And: when changes? in which formats/systems? with/without termbase!


Key performance indicators: costs
For a specific application scenario:

 Procurement costs for a termbase system (12,000-40,000 €)

 Costs for system support and update (1,100-9,000 €)

 Time for one SL term entry (ø 30 min) plus TL info (ø 20 min)

 Number of new entries/year (60-600) + updated entries (100-1200)

 Monthly salaries for the terminology staff


Cost-benefit analysis
Type of Problems
Degree of Necessity
Degree of Effects
e.g. Number of Languages

Framework Conditions e.g. Amount of Translation
for Optimum Usage e.g. TMS Usage
e.g. Number Employees TD

Benefit Key Indicators: Cost Key Indicators:
e.g. Costs for Source Text Changes e.g. Investment Costs for TermBase System
Costs for Answering Translators Questions Costs for Training Personnel
Costs for Target Text Corrections Running Costs for Terminology Management
TMS Match Rates System Maintenance

Alternatives
Without Defined With Defined (Without Defined With Defined
Terminology Terminology Terminology) Terminology

Figures to Compare from Benchmarking or Estimation

Evaluation and Comparison Benefit Evaluation and Comparison Costs

Cost-benefit analysis: sample
+ 200 000 € 1. Year 2. Year 3. Year 4. Year

+ 150 000 € Corrections Corrections Corrections
Break-
Even-
Translation Translation Point Translation
+ 100 000 € costs costs costs

+ 50 000 € Translator Translator Translator
queries queries queries

Changes Changes Changes

- 50 000 € Salary Salary Salary Salary
expenses expenses expenses expenses

- 100 000 € Licenses ROI Licenses Licenses Licenses
Initial
investment

- 150 000 €

- 200 000 €


Pain curve for terminology management

Source: Dunne, Keiran; Multilingual, April 2006

Terminology error propagation

Source: Dunne, Keiran;
Multilingual, April 2006


Source: Childress, Mark;
Multilingual, April 2006

Terminology error propagation


Terminology extraction
 In many application scenarios of terminology work, the extraction of
terminology from (existing) textual material is recommended.

 We can differentiate between the following extraction methods:

 Monolingual term extraction (text in electronic form)
 Bilingual term extraction (parallel aligned texts, i.e. TMs)
 Manual (human) term extraction
 Computer-assisted term extraction (tools propose term candidates)
• With statistical methods (for “all” languages, cannot use
knowledge about syntax)
• With linguistic methods (better results, but only for “important”
languages)
• With hybrid methods (combining statistical and linguistic methods)

Terminology extraction tools
 Features of (monolingual) term extraction tools:
 Common functionalities from concordance programs (e.g.
WordSmith): identify words, word statistics, KWIC index,
alphabetic/frequency order
 Reducing inflected word forms to the basic canonical form:
needed for real statistics, needs morphological knowledge
 Filtering and ignoring function words (articles, conjunctions etc.)
and general language words (but what is general language?)
 Filtering and ignoring terms that are already included in a term
base
 Identifying multi-word terms, noun phases and verbal phases
 Identifying discontinuous elements and elliptical constructions


Improving and
enriching term
candidates with
SDL MultiTerm
Extract



Settings to
improve the
results of term
extraction with
SDL MultiTerm
Extract



 Benefits and problems of term extraction tools:

 Term extraction tools are helpful in preparing terminology for
large translation projects (with several translators) and for an
initial feeding of a term base (with company or subject specific
terminology)

 Result of a term extraction is a list of term candidates; the list
must be checked; but what about the texts (with possible not
extracted terms)?

 Results are only terms (and context examples), but no other
terminological information; it is a kind of a to-do list for the
terminologist

 The more linguistics the better the results; but what about “less
common” and minority languages? K.-D. Schmitz, IIM, FH Köln

Terminology management systems

 Terminology management systems are software
applications that are designed to manage terminological
data.
 They support tasks related to terminology work and store the
results: Terminological data can be entered, edited, deleted,
retrieved and filtered.
 Most of the systems available on the market are based on
(relational) data base systems (MS-Access, SQL, Oracle).
 Can be seen as a kind of CAT-Tools (CAT=computer assisted
translation).
 Tables in word processing or spreadsheet programs are not
adequate for terminology management !


Terminology management systems

Classification of terminology management systems:

 Complexity (languages): monolingual / bilingual / multilingual
 Entry structure: predefined / free-definable / hybrid
 Autonomy: autonomous / CAT tool component / hybrid
 Software technology: stand-alone / client-server / browser-based
 Business aspects: proprietary / commercial / open source

e.g. SDL MultiTerm 2009


Designing a terminology management solution

Before designing a terminology management
solution and choosing, adapting or programming a
terminology management application:

 Analyze the needs and objectives
 Specify the user groups, tasks and workflow
 Define the terminological data categories needed
 Take into account the basic modelling principles
 Model the terminological entry
 Select, adapt, develop the software
Meta data

Typology of data categories I
 Complex data categories
 Open data categories
content not predictable and defined by specification
e.g.: term, definition, note
 Closed data categories
content defined by a limited set of possible values
e.g.: gender, part of speech, geographical usage

 Simple data categories
content only yes or no; values of closed data categories
e.g.: masculine, noun, DE


Typology of data categories II
 Concept-oriented data categories
e.g.: subject field, figure
 Language-oriented data categories
e.g.: definition ?
 Term-oriented data categories
e.g.: part of speech, context
 Administrative data categories
e.g.: author, date, note
 Special data categories
e.g.: term, language, (structural elements), (shared resources)


Concept orientation

 All terminological information belonging
to one concept including all terms in all
languages and all term-related and
administrative data must be store in
one terminological entry

concept = terminological entry


Lexicographical view / model / entry

meaning meaning
word

meaning meanding

meaning meaning


Terminological view / model / entry

term concept term

term term

term term
descriptive
terminology management

Terminological view / model / entry

term concept term

(term) term

term term
prescriptive
terminology management

Lexicographical entry


Terminological entry


Eintragsmodellierung + Prinzipien


Term autonomy

 All terms belonging to one concept should
be managed (in one terminological entry)
as autonomous (repeatable) blocks of data
categories without any preference for a
specific term

 Therefore all terms can be documented with the
relevant term-related data categories
 Term autonomy is necessary for the main term, all
synonyms, all variants, and all short forms
 Term autonomy is not explicitly discussed in
theoretical literature

Concept orientation & term autonomy
TermEntry
Concept
represented by ID-No. and/or classification / notation

Language 1 Language 2 Language 3 ...
+ AuxInfo + AuxInfo + AuxInfo
Term 1 Term 1 Term 1
+ AuxInfo + AuxInfo + AuxInfo

Term 2 Term 2
+ AuxInfo + AuxInfo

Term 3
+ AuxInfo


Concept orientation & term autonomy


Terminological data modeling

 Terminological data categories
 ISO 12620:1999 / 12620:2009
 definition, subject field, grammar, context,
project code, author, date etc.
 Data Category Registry (DCR)

 Terminological data modeling principles
 ISO 12200 / 16642 / 30042
 meta model, concept orientation, term autonomy,
TBX (Termbase eXchange)

Terminology control

 In many application scenarios of terminology work, the checking of
correct and consistent use of terminology in documents (created by
technical writers or translators) is recommended.

 We can differentiate between the following control methods:

 Monolingual terminology control
 Bilingual terminology control (for translations)
 Manual (human) terminology control (part of proof reading & QA)
 Computer-assisted term control (tools analyze and check documents)
• Without linguistic methods (for “all” languages, using the content
of a term base)
• With linguistic methods (better results, but only for “important”
languages)

Terminology control tools

 Features of terminology checking tools:
 Similar to spell checkers and auto correction

 Integrated into editors, authoring systems, CAT tools, but also as
stand-alone programs
 Directly during the writing process of a document or translation, or
as an autonomous process (when the document is finished)
 Connection to the term base entries (interactive or via
export/import)
 Very often combined with grammar and style checking (controlled
language)
 Using fuzzy search and/or linguistics (inflected terms in texts vs.
canonical form of the terms in term bases)
 Deprecated terms must be maintained in the term base (no-terms)



Sample of a terminology check with acrolinx IQ Suite


Automatic detection of
linguistic variants with
acrolinx IQ Suite


Conclusion 1
 You will find terminology problems in many companies
 In any case, terminology management will improve
important factors such as:
better communication, consistent corporate language, avoidance of
errors and misunderstanding, faster training, better translations

 Not all benefits are easy to calculate and quantify
 Benefits and ROI depend on several company specific
framework conditions
 Terminology management is no luxury, it is a necessity
for all industrial companies and service providers,
and this can be documented by key indicators

Conclusion 2
 Only excellent managed terminology can guarantee
a high quality information and communication:

 follow guidelines for term creation and
term selection

 model your termbase with concept orientation
and term autonomy, and include appropriate
data categories for documenting terms

 Wrong decisions and mistakes can later be
repaired only with huge efforts and costs !

Thank you for your attention

Prof. Dr. Klaus-Dirk Schmitz
Fachhochschule Köln
Fakultät 03 - ITMK/IIM
Mainzer Str. 5
50678 Köln
klaus.schmitz@fh-koeln.de

Prof Klaus: Terminology Management

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (12)

Similar to Prof Klaus: Terminology Management

Similar to Prof Klaus: Terminology Management (20)

More from akashjd

More from akashjd (8)

Prof Klaus: Terminology Management