2. What Is TERMinology?
The system/science of Terms
TERM: Silicon Valley (EN source)
Силиконовая Долина (RU target)
िसिलिलिकॉन वैलिी (HI target)
Term (Latin: terminus — limit, border)
a word or group of words designating something,
especially in a particular field
or
smth. to be translated consistently throughout the text
http://dictionary.reference.com
3. We All Use Terminology
in everyday
life and work
8. Glossary Development
must go in Parallel with
Documentation Writing
Content=Documentation+Unified Terminology Storage
Graphics: FLDSMDFR (“Cloudy with a chance of meatballs”) from youtube.com
9. Unified Terminology
Storage & Management
Benefits & cost reduction
• Independence from CAT-tools. No need to
– train resources for each new TM tool
– convert glossaries between CATs
Software developer
perspective
• Easy customers’ involvement in localization
LSP perspective
More relationships with customer
Prompt understanding and approving => less
terminology corrections => save Time & Money
• Product Community involvement
User perspective
It’s hardly possible to Integrate Users into CAT,
but it’s easy in unified terminology storage
10. TYPICAL ILLUSIONS: False Friends
No glossary => No pain
No glossary => More pain:
Without a Terminology Management System (TMS)
• Terminology is hardly controlled
• Translation is inconsistent
• Errors are easily propagated and hardly found
In localization projects each term should be:
Additional support point
Low cost checkpoint
11. Term Mining
Usual ways
•Term extraction utilities, based on statistical methods
•Manual gathering from bilingual content at the localization stage
•Ideally: in parallel with technical writing
Usual problem
Reducing costs on term mining
12. Amount of terms, added by user.
Key Performance Indicators
Example. No terminology on significant project.
No time and/or budget for term mining
Localization-without-terminology
is a Minefield
On-Click Add to
the Glossary!
Give the team a chance!
14. Smart Term Mining Results
The Glossary you gathered and saved in the
Unified Storage
is your Intellectual Asset
which can be successfully reused
because it is independent from CAT and other
formats.
15. How do we manage gathered terminology
in lots of different formats?
Software developer
Data to localize:
Data to localize:
.json, .xml, .js,
.json, .xml, .js,
RC, .yml, etc.
RC, .yml, etc.
Context Data:
Context Data:
PDFs, HTML, etc.
PDFs, HTML, etc.
Bilingual CAT
formats
.xliff, .ttx, etc.
Localization
Vendor
Independent
Testers
and
Reviewers
Loc. Builds
Loc. Builds
Data
17. Case Study. Effective Environment for
Successful Terminology Management
Software developer
Bilingual CAT
formats
.xliff, .ttx, etc.
Data to localize:
Data to localize:
.json, .xml, .js,
.json, .xml, .js,
RC, .yml, etc.
RC, .yml, etc.
Localization
Vendor
Terminology
Terminology
Storage & QA
Storage & QA
Infrastructure
Infrastructure
Context Data:
Context Data:
PDFs, HTML, etc.
PDFs, HTML, etc.
Independent
Testers
and
Reviewers
Loc. Builds
Loc. Builds
Data
Terminology
18. Over-managed Term Management Systems
LSP example:
No possibility to add a term directly to client’s TMS =>
We sent a request as per instruction and waited.
3 months passed, no term was added.
19. TYPICAL ILLUSIONS: False Friends
My Terminology Management Tool already supports all languages.
Enough for proper terminology control
In flectional languages
Linguistic Support in TMS must not mean only storage,
but associated Quality Assurance procedures
Choose the tool with minimum level of false
positives in terminology check!
20. Case study. Fast Terminology Control.
Be able to check quality level of linguistic
services!
Example. Brilliant glossary on the project.
Are terms translated properly throughout the translation?
Unified TMS to easily interact with QA-tool
to build a terminology check report on the selected glossary.
The same applied to
• Tags
• Consistency (duplicates,
reverse duplicates)
• Spelling
• Numbers
21. Language Specifics: Morphology
• Lots of false positives
• Multipage QA-reports
TMS Morphological control
must vary depending on the language
Chinese: comparing 2-3 sequential bytes in glossary
with the same bytes in translation
Russian: up to 12 variants of the same word possible
Punjabi: 15-20?
22. Case study. Be language-scalable!
Use different language pairs and projects in TMS
Example. Multilingual project. Lack of terms in one of the languages in
TMS (new team involved).
Give the team a chance to utilize examples of other languages!
Multilingual glossary with concepts for users to consider the
meanings of the terms
23. Formulate and Prioritize Your Needs
Include in
Term Management System
Source and Target
Context
Concept
Term Confidence Level
Integration with QA
Different language pairs and projects
24. Formulate and Prioritize Your Needs
Exclude from
Term Management System
Learning the glossary by editors and terminologists
Manual notifications on changes in Excel, Google-docs
Terminological plug-ins
Generalities
25. Illustrated Problems and Solutions.
How to:
avoid delays in preparing glossaries for
localization of your product
check the work of language services provider
increase quality of the localized product
save time and money
Unified Terminology
Storage
and Management!
26. SUMMARY.
PROPER TERMINOLOGY MANAGEMENT SYSTEM
TO MAKE YOUR TERMINOLOGY FIT!
The work of the person,
who makes terminological decisions
—
The thing you can’t exclude from
terminology management
Terms live their own life, from birth to
death, and you need to treat them
properly to make them fit and working
Every software developer is writing the Documentation.
If there’s no glossary involved, we as LSP localize just tons of technical materials, constantly wondering “what this device actually is?”
But if technical writers already have a unified terminology storage, they also have a powerful tool to make the terminology unified throughout the product.
Let’s see, how.
First of all, by Unified Terminology Storage we don’t mean Excel-files or google-docs being exchanged among the team.
You get all the team involved in the process, providing the access to the terminology portal, which means a chance to store, add, update and control terminology on every project stage.
Attention! Bilingual formats need to be constantly accessible to all of the participants of the process! Without bilingual files no automated Quality Checks are possible at all… Neither is Terminology control!
Note! Unified Terminology Management & QA is separated from the CAT! Choosing the built-in services will only add costs, as all the teams would need to install and examine them (remember slide 9?).
As LSP we had a client with own TMS and we couldn’t really interact with it.
Integration & Interaction Issues: What is the average time of manual term retrieve?
Don’t fall into the same trap and see what others invented.
May be, taking into account the growing variety of Terminology Management systems, the best solution is using required term management options and facilities through API.
Illusion typical for both creators of content and localization teams.
What does “support” mean for terminology system?
Have you just got UNICODE support in your system?
Is that enough for your language sets?
How can your system provide checks for terminology consistency?
Who will correct terminology inconsistency?
Having a QA-report you easily see, how many and severe the translation errors are.
But sometimes you still need a person with a linguistic knowledge to evaluate or correct the errors.
But speaking of linguistic quality checks, we need to remember that strategy of morphological control is different for different languages.
Simple QA-report is not enough for inflectional languages.
Morphology leads to lots of false positives in multipage QA-reports.
When making the decision of purchasing the terminology portal/storage, at first you perform a (little) investigation of your needs.
The most effective way to make your terminology fit is
the right choice of the Tool or TMS, which covers all of your terminological requirements.
The talk on terminology must be much longer, especially taking into consideration the growing variety of corresponding tools.
We illustrated only several effective ways of struggling through everyday terminology challenges, which in the process of this presentation we also have tried to analyze a bit.
And if you ask, why comparing them to problems of fitness, now, you may guess that the main reason is:
Terms live their own life, from birth (meaning offered term status) to death (meaning obsolete), and you need to know how to treat them properly to make them fit and working.