TAUS Quality Dashboard

From Quality Evaluation to Business Intelligence

March, 2016

1
About this document

This document describes how the TAUS Dynamic Quality Framework (DQF) generates a Quality
Dashboard serving the interests of all stakeholders in the global translation industry. Through
easy-to-use plug-ins translators and managers share data and reports that give them valuable
statistics, benchmarking and analytics.
The Quality Dashboard is developed for and with the TAUS members and users.

COPYRIGHT TAUS 2016

All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system of any nature, or
transmitted or made available in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without the prior written permission of TAUS. TAUS will
pursue copyright infringements.
In spite of careful preparation and editing, this publication may contain errors and
imperfections. Authors, editors, and TAUS do not accept responsibility for the consequences
that may result thereof.
Published by TAUS BV, De Rijp, The Netherlands
For further information, please email info@taus.net

2
Table of Contents

1. Background 4
2. TAUS Industry Program 7

3. The Principles Underlying the TAUS Quality Dashboard 8

4. Benefits 10
5. From Quality Evaluation to Business Intelligence 11
6. What is the Quality Dashboard 13
7. Implementing the Quality Dashboard 18
8. How to Get Started 20
9. Appendix 21

3
1. Background

Since it was founded in 2005, TAUS has always followed the industry trends, providing
guidance to the stakeholders in the translation industry. Originally started as a Think Tank,
TAUS has gradually expanded its services to the general language service industry, first
focusing on automated translation and in the last five years on quality evaluation. TAUS is
currently working on the development of the Quality Dashboard an industry-shared platform
for measuring productivity and quality and turn data into business intelligence. This section
briefly outlines the evolution that TAUS has undergone in the past few years, focusing on
quality evaluation, and highlights the main achievements and milestones up to the present
days.

2011
Old school: one size fits all
Translation quality evaluation is problematic. In 2011, TAUS conducted a survey among its
members. We found that despite very detailed and strict error-based evaluation models the
satisfaction levels with both translation quality and the evaluation process itself were very
low. QE models are static, that is, there is a one size fits all approach. Little consideration is
given to multiple variables such as content type, communications function, end user
requirements, context, perishability, or mode of translation generation (whether the
translation is created by a qualified human translator, unqualified volunteer, a machine
translation system or a combination of these.)

Dynamic Quality Framework

In November 2011, TAUS published the foundational report for the Dynamic Quality
Framework (DQF). (Download this report.) The solution proposed in this report was a dynamic
evaluation model that takes into account the changing landscape of diversification in content
types and the adoption of automated translation technologies.

Content Profiling
In collaboration with members and researchers from Dublin City University, TAUS designed a
model for profiling content types based on three vectors: utility, timeliness and sentiment
(UTS). The UTS scoring started to give users guidance on methods for translation quality
evaluation.

2012
DQF Knowledge Base
In 2012, TAUS started developing a Knowledge Base for everyone working in the global
translation industry documenting different approaches to translation quality evaluation.

4
The models described in the knowledge base are: adequacy (accuracy), fluency (readability),
usability, community feedback, error typology, productivity measurement. The knowledge
base contains use cases, templates, metrics and specifications. In consultation with users and
members, TAUS developed best practices for post-editing, sampling and different quality
evaluation techniques that are all part of the knowledge base. The knowledge base can be
accessed from the TAUS Academy section of the website.

2013
DQF Error Typology
A vast majority of the providers and buyers of translation services manage their quality
program with an error typology template. The LISA QA model which was developed in the
nineteen-eighties forms the basis of the error categories being applied in most cases. The SAE
J2450 is another metric that is well-known in automotive translations. TAUS worked with Dr.
Sharon OBrien to develop a more up-to-date version of these error typologies and made it
available under DQF. In 2014, DFKI published the MQM (Multidimensional Quality Metrics).
Under the European funded project QT21, TAUS and DFKI have harmonized the DQF and MQM
error typologies.

DQF training program

In 2013, the Localization Institute started offering a quality management training program
featuring DQF. In cooperation with TAUS, Willem Stoeller has now developed a much improved
and more focused DQF training program is offered online through the TAUS web site. Through
this course, industry professionals will be trained in the DQF methods and best practices
ensuring a consistent usage of the evaluation tools.

2014
DQF Tools
On request of its members, TAUS developed various tools for quality evaluation and made
them available through the web site. Since early 2014, users can go to https://dqf.taus.net/
to do a number of quality evaluations: adequacy, fluency, error review, productivity
measurement, MT ranking and MT comparison.

Free access to DQF for students

Since December 2014, TAUS offers translation students at universities around the world access
to all TAUS Data and resources, including the DQF knowledge base and tools through the Free
Academic Membership Program.

2015
Quality Dashboard and DQF API
After the adoption of DQF Tools by over a 100 companies, users expressed interest in
accessing DQF from within their translation technology of choice. Therefore, TAUS set forth in
developing an API that would enable users to connect directly to DQF. More and more data
were sent to DQF which required a way for users to access them. This is how the Quality
Dashboard came into existence. The Quality Dashboard represents the platform where users
can access their data and gain business intelligence to further improve their processes.

DQF Users and User Group

An active user group has formed around DQF, allowing the TAUS team to constantly improve
and fine-tune their services. An early user group centered around the DQF tools, which were
designed for incidental quality evaluations or for the comparison of the quality of MT engines.

5
After developing the TAUS API and a number of integrations of DQF into existing translation
technologies, a new community has been formed around DQF, bringing together all TAUS DQF
users to share best practices, give feedback and put forward suggestions for further
developments. The DQF User Group can be found on LinkedIn.

DQF as industry reference

Quality evaluation is one of the most pressing themes in the translation industry. Operators
realize that, unless they start relying on common metrics and using the same methods and
tools, they run the risk of comparing apples with oranges and taking uninformed decisions.
DQF is more and more referenced as an industry standard and best practice for translation
quality evaluation.

6
2. TAUS Industry Program

TAUS is a think tank and platform for industry-shared services and resources for the global
translation sector. The TAUS industry program is focused in three main areas:

The Quality Dashboard is an industry platform for statistics on translation, benchmarking

translation activity and quality and analysis of translation performance and production. TAUS
provides API specifications allowing translation technology providers and users of translation
services to plug TAUS DQF into their work environment.
The TAUS Academy is a knowledge center where users can find training courses for post-
editing and quality management with DQF, a library of reports and articles on translation
automation and quality evaluation, industry best practices, use cases, technology briefings
and webinars. The TAUS Academy also contains directories of translation technologies,
certified post-editors and profiles of member organizations.
The Data Cloud is the largest industry-shared repository of translation data, containing more
than 70 billion words in 2,500 language pairs. Users can download and pool data for the
training and customization of translation technologies. Members can choose to share data
publicly or limit the sharing in a Private Vault. TAUS provides an API allowing the integration
of the Search and Data Pooling functions in other translation work environments.
This document is about the TAUS Quality Dashboard. Translators, project and vendor managers
as well as buyers of translation who are interested in using the TAUS Quality Dashboard will
find all the relevant information in this document. Developers, interested in using the plug-in
to integrate their technology with the TAUS Quality Dashboard can find the API specifications
on the TAUS website.

7
3. The Principles Underlying the TAUS Quality Dashboard

The TAUS DQF and Quality Dashboard have been developed with some important principles in
mind. These principles guide the current implementation of the Quality Dashboard and help
drive the future developments and improvements of the TAUS services.

1. Industry-shared metrics
Translation quality is a pressing theme in the translation industry. The diversification in
content types and rapid adoption of translation technologies (including machine translation)
drives the need for more dynamic and reliable methods and measurements of translation
quality evaluation. The vision behind DQF is to standardize the methods and tools of quality
evaluation and to aggregate the scores and measurements and make these available through
industry-shared metrics. These new metrics will promote standardization within the industry
in the area of translation quality evaluation.

2. QE as business intelligence
Quality evaluation is costly and troublesome in many translation environments today.
Measurements are often subjective and anecdotal. Industry-shared metrics will lead to more
reliable measurements that give translation operators and producers (translators) useful
insights that help them to adjust and improve processes. Today, for most companies quality
evaluation is a necessity without remedy. The industry-shared metrics will turn quality
evaluation into business intelligence steering and supporting management decisions.

3. Benchmarking
Translation is often perceived as a black box. It is not always clear to the customer what is
involved in the production of a translation. Language service providers apply different
methods, tools, processes and metrics to assess the quality of translation, which makes it
impossible for buyers of translation to make useful comparisons between services and
technologies. Adoption of DQF will lead to standardization of metrics and processes. This
standardization will allow buyers of translation then to compare and benchmark the
productivity and quality of translation, both across their own translation resources, projects
and language resources (internal benchmarking) as well as against peers in the industry
(external benchmarking).

4. Democratic
The Quality Dashboard promotes democratic access to data. Every stakeholder in the supply
chain - the procurement officer, vendor manager, project manager, quality manager,
translator - can get access to valuable and relevant reports on the Quality Dashboard. Access
is regulated by user profiles making sure that everyones access is limited to the reports that
s/he is authorized to view.

5. Community-Driven Roadmap
Just like the original project of the DQF Tools, the Quality Dashboard has been developed
following the needs and requirements of the stakeholders in the industry. TAUS constantly
collects feedback and input from its members and users to determine the priorities for
feature development.

8
6. Open Source API
The key to drive industry-wide adoption of the Quality Dashboard is the API that connects the
translation workflows and translation tools with DQF: the standard Dynamic Quality
Framework. This DQF API defines the agreed metrics - like the harmonized DQF-MQM error
typology - and specifies the data points. TAUS is making the DQF API available under an open
source license. Every translation operator and technology developer has free access to the API
and can easily integrate DQF in their workflow and translation tool.

7. Measure all translation output

The Quality Dashboard is designed to report on the productivity and quality of all translation
output, not just of machine translated content.

9
4. Benefits

TAUS DQF and the Quality Dashboard have been developed with the whole industry in mind,
so that every stakeholder in the language services industry can benefit from adopting and
using the DQF integration. This section briefly summarizes the main benefits for DQF users.

Industry benefits
DQF is an industry collaborative program that elevates the discussion about quality to an
industry level. On an industry scale, using DQF helps towards:
1. Increased credibility of the quality assurance function and process throughout the
industry. This will improve the image of the industry in general and attract more
customers, allowing the translation sector to grow faster.
2. Lower cost in general for all translation operators as a result of the shared investments in
resources and tools.
3. Increased satisfaction with both vendors and clients by referring to standard metrics.

Business benefits
Users of DQF can benefit from this industry collaborative program in multiple ways:
1. Sharing knowledge and standard references. Stepping away from the rigid outdated error
typology review as the only way to measure translation quality, each individual translation
operator does not have to research new methods on his own. Referring to common
methods and best practices avoids confusion and disputes and ensures consistency.
2. Reduced costs. Sharing resources, tools and knowledge allows each individual stakeholder
to reduce the costs of overhead in quality research and tool development.
3. Increased efficiency. Agreed best practices and shared tools and metrics reduces friction
in each client-vendor relationship and makes it easier to collaborate in a multi-vendor
environment.
4. Increased flexibility. Industry consensus and benchmarking on quality evaluation methods
and scores allow for a more flexible use of often more efficient and less costly methods of
quality evaluation.

Professional benefits
While the adoption of DQF may start at larger companies, individual translation professionals
are also included in the vision behind the Quality Dashboard. In particular, they can:
1. Benchmark their performance against industry averages to assess their translation
throughput.
2. Benchmark their performance against their previous projects to track their improvements
both with the aid of specific translation tools and without.
Keep track of their performance over time and having an automatic track history of projects
which they simply executed for a customer or even create themselves whenever DQF is not
already in use.

10
5. From Quality Evaluation to Business Intelligence

The DQF tools were initially designed for incidental quality evaluations of MT projects, for
instance for tracking and benchmarking of translators or translation vendors from time to
time, or for the comparison of the quality of MT engines. Following the requests from many
users, TAUS has developed an API that allows users to make use of DQF while in their
translation production mode.
The main function of the DQF integration is to enable process optimization, particularly when
it comes to quality evaluation. Most industry sectors are focused on increasing the quality of
their services, products and processes. For most buyers of translation and even the providers
and the translators themselves, translation is like a black box. They miss the insights, the
data and the business intelligence to take well-grounded decisions about processes, tools,
resources, budgets.
The Quality Dashboard will help all stakeholders getting the indispensable statistics,
benchmarking and analytics to take well-informed decisions. In this chapter, we list examples
of the questions that different stakeholders may have and that the Quality Dashboard will be
able to answer. Some examples of business-related questions from each stakeholders group
are listed below.

Buyers questions
1. How are we performing compared to industry peers, in terms of production throughput
and quality?

The Quality Dashboard provides the benchmarks by several categories such as language,
content type, industry.
2. Should we consider using MT for support content?

The Quality Dashboard allows to filter data per process (e.g. MT) and per content (e.g.
support content) and also shows the average productivity and quality scores.
3. How is our translation technology ranking compared to industry?

The Quality Dashboard shows the scores for different technologies, also broken down by
language, industry, content types.
4. How good is our MT engine?

The Quality Dashboard will show the productivity and quality scores for a given engine/
technology.
5. Are we paying our vendor/translators fairly?

The Quality Dashboard shows the average productivity scores (number of words per hour)
for vendors/translators and for the specific vendors employed for a given project.
6. How does the quality compare between our languages and vendors?

11
 The Quality Dashboard shows quality scores by the categories selected by the project
manager or quality manager.
7. How efficient are own internal reviewers?

The Quality Dashboard not only tracks efficiency of translators and post-editors but also of
reviewers and editors.

LSPs questions
1. What is the correlation between the productivity (number of words per hour) and the
number of edits per hour, on total and per translator?

The Quality Dashboard provides the numbers and the benchmarks by vendor, language,
content type, industry.
2. How did our scores change since we changed our translation technology platform?

The Quality Dashboard shows the productivity and quality scores by any selected time
period. It will also enable a timeline mode to monitor the evolution of performance over
time.
3. Should we change to a new technology for the translation of our web content?

The Quality Dashboard shows the productivity and quality scores by any selected
technology and content type.
4. Where did our translators source their translated segment from?

The Quality Dashboard shows distribution of segments and presents the percentage of
human translation, machine translation, translation memory, post-editing.

Translators questions
1. How is my productivity compared to industry averages?

The Quality Dashboard provides the numbers and the benchmarks language, content type,
industry.
2. How am I doing on this project compared to that project?

The Quality Dashboard shows the productivity and quality scores by any project and also
allows filtering by customer.
3. Am I using the best tool for the job?

The Quality Dashboard shows the productivity and quality scores by technology in
comparison to other technologies so that translators can possibly add more efficient tools
to their translation toolbox.

12
6. What is the Quality Dashboard

The Quality Dashboard delivers on the DQF vision described here above. It is an industry
collaborative platform for the global translation services sector, allowing every stakeholder
within the supply chain to access the reports that are relevant for their activity. There are
different levels of stakeholders in a standard supply chain. For example, in a supply chain
consisting of four levels, stakeholders are the Buyer, the Language Service Providers (LSPs)
and the Translators/Reviewers. In a four-level chain, we further distinguish between larger
LSPs (Multi-Language Vendors, MLVs) and smaller LSPs (Single-Language Vendors, SLVs). The
lines represent the connections between the different stakeholders, the dotted lines
represent other possible, though less likely, connections.

The Quality Dashboard collects and displays data about productivity and quality for all
projects and users across the supply chain. In other words, all the people involved in a
project will be able to access the reports relevant for their activity. This will happen on the
Quality Dashboard interface, where users will find productivity and quality metrics across
content types, industries, processes used, technologies applied and by language pairs, among
others.
The Quality Dashboard can be accessed by anyone using the URL: https://dqf.taus.net/
quality-dashboard

13
1. Productivity

The landing page of the Quality Dashboard displays the average productivity score across the
whole industry in terms of words processed per hour as well as the average edit rate score.
Users can then navigate through the available filters using the menu on the left hand-side.
Each filter allows to highlight a specific parameter or property of the project such as the
language pair, the process used, the technology involved and so forth.
These reports allow users to compare the industry averages for each of these more specific
filters in terms of productivity.

Productivity is the throughput or speed expressed in the number of words per hour. This is
currently the most common way of measuring translation performance. Productivity is an
absolute score. The distinctive feature of the Quality Dashboard is the possibility to measure
ones own performance against these average values. In order to access the benchmarking
feature, users need to be logged in and need to have used DQF for their projects.

14
These project-based attributes will be complemented by filters to track the performance at
the customer and vendor level as well. These filters will enable each stakeholder to access
the relevant information for their role in the supply chain. This means that a translator will
only be able to access data about his/her own performance with respect to the customer(s)
he/she is working for.

This data will be aggregated per customer but additional filters can apply.
Similarly, an LSP can monitor its own performance towards the companies they are working
for. Because an LSP also works with vendors, they will be interested to see how their vendors
performed for a specific project or aggregating data for a specific language pair. For an LSP
the tracking goes both directions: towards the customers and towards the vendors.

15
2. Quality
The Quality reports of the Dashboard include similar filters to the ones under productivity.
The Quality section provides stakeholders with quality-related information on their project. A
quality review can consist of two main activities: flagging errors in a segment and/or
correcting them. TAUS has adopted an industry-standard error typology which is the DQF-MQM
harmonized model. This model is the result of a cooperation between TAUS and the DFKI
institute who worked on developing the MQM error typology as a deliverable in the EU project
QT21. MQM (Multidimensional Quality Metrics) is a framework for building task-specific
translation metrics. It allows users to create custom metrics that can be used for various
assessment purposes. By providing a master vocabulary of error types, users can describe
metrics in a fully transparent fashion. MQM has been implemented in a variety of commercial
and open-source tools and now the harmonized model DQF-MQM is available for quality
measurement.
For a detailed overview of the error categories in the DQF-MQM harmonized error typology
and severity levels, please refer to the Appendix.

Under the quality sections, users can have an overview of the number and type of errors
found in translation. Every error is also associated with a severity level which determines the
weight each error has in the text. Based on the final score a pass/fail threshold can be
determined. Similarly to the productivity reports, quality reports will also offer a measure of
the changes made to the translation in the form of Correction Distance.

16
The Quality Dashboard is developed for and with TAUS members and subscribers. QE Summits
are organized to bring users together and agree on best practices and new features and
support for the Quality Dashboard. They take place twice a year, hosted by TAUS member
organizations. TAUS also organizes bi-monthly user group meetings and webinars for new
users. On the TAUS website, users can benefit from an extensive knowledge base and
community discussion forum.
The Quality Dashboard will follow the needs of the industry and continuously add new tools
and features. The release of a Quality Control environment (based on the harmonized DQF
and MQM error categorization) with the possibility for companies to do error annotation and
correction is on the roadmap for 2016. There will also be an API for the Quality Control tool to
allow members and subscribers to plug this service into their common translation
management environment. Additional new features will be announced in the course of the
year.

17
7. Implementing the Quality Dashboard

The Quality Dashboard and its reporting rely on the data supplied by users of translation
technologies. Each technology can be linked to the Quality Dashboard by means of the TAUS
API. This API is developed so that the required data can be collected from the translation
system and sent to the DQF server for the reporting part. The API is released under and Open
Source license so that anyone can use it to connect their tool to the TAUS DQF.
API specifications are available on the TAUS website and additional support for the integrators
will be provided directly by the TAUS team or TAUS partners. If you have any integration-
related questions, please write to dqfapi@taus.net. There is also an open GitHub repository
available for developers. Just contact TAUS to be added to the team.
Once data flow into the DQF Server, the system will recalculate the industry average scores to
be displayed on the Quality Dashboard landing pages. These reports are accessible to anyone
visiting the Quality Dashboard. However, for more advanced reporting capabilities users need
to:
- ensure that their translation technology of choice is properly connected to the DQF API
across the project supply chain
- sign up for a TAUS account in order to access more reports on the Quality Dashboard
A TAUS account is required for every user interested in accessing the reports. An account will
also provide them with credentials to use their translation technology in combination with the
TAUS DQF. The TAUS account can be of two types: a subscription or a membership.
A user-subscriber is generally an individual working as a freelancer or a small-sized company.
The first subscription level (called Basic) is free and gives access to a limited number of
reports which only include industry average scores.
The professional subscription level enables individual users to benchmark their performance
on a number of reports. This subscription level is meant for translators, consultants and
individual users.
With the Business subscription, companies will have access to the full content of the Quality
Dashboard both in terms of reporting and benchmarking.
The paid subscription plans give users access to additional TAUS resources such as DQF Tools,
TAUS Data Cloud and reports.
The TAUS Membership offers companies additional benefits compared to the subscription
plans, such as discounts for TAUS events and partner events, increased visibility for the
company and a more advantageous pooling ratio for the TAUS Data Cloud. The TAUS
Membership fee is calculated on the number of user log-ins required. The Basic level comes
with 1 user account, but can be upgraded by the Member, depending on how many user log-
ins they need.
The subscription plans and the TAUS Membership will determine the types of reports a user is
entitled to see and use. It is useful to remember that the reports a user will see via their paid
account will still depend on their respective role within the supply chain, as described in the
previous section. For example, a small SLV with a business account will see the reports

18
relating to the companys activity, i.e. the company performance towards their customers
(buyers or larger LSPs) and their vendors (translators). Similarly, two translators will only get
to access data about their own performance towards their customers but if they have two
different subscription levels (i.e. basic and professional), the former will have a more limited
access to reports than the other translator with a professional subscription.

19
8. How to Get Started

There are several ways to get started using the TAUS DQF and the Quality Dashboard. Please
see below to learn how to generate reports in the TAUS Quality Dashboard or integrate the
TAUS DQF API in your translation technology.

For Users
You can access the TAUS Quality Dashboard here and obtain free reports with industry average
data on productivity. To see more advanced and personalized reports and benchmark your
own data against industry average, you need to have a TAUS account. If you do not have an
account with TAUS yet, you can subscribe here.
Visit the DQF users page to find your translation tool(s) of choice and follow the instructions
on how to get started. Once you start using DQF for your projects, the TAUS Quality
Dashboard will generate productivity and quality reports for you.
Should your tool(s) not be listed, please let us know. Our Member Services team is always
available to help you and answer any questions you may have at memberservices@taus.net.

For Developers
TAUS welcomes any translation technology provider who would like to integrate DQF into their
CAT tool or Translation Management System. You can obtain more detailed instructions on how
to become an integrator in the "For Developers" section of the TAUS website.
You can obtain the free TAUS DQF API specifications directly from the TAUS website or from
Github. Please write an email to dqfapi@taus.net if you have any questions on how to get
started. We are happy to welcome you to the developers community.

20
9. Appendix

The appendix provides a detailed descrip2on of each error category of the harmonized DQF-MQM error-
typology including deni2ons and examples. The second table explains in greater detail the dierent
severity levels that can be applied to each error category.

Harmonized DQF-MQM Error Typology

High-level
error Granular
ID types error-types Deni7on Example

The target text does not accurately

reect the source text, allowing for any Transla2ng the Italian word 'canali' into English as 'canals'
1 Accuracy dierences authorized by specica2ons. instead of 'channels'.

The target text includes text not present A transla2on includes por2ons of another transla2on that
11 Addi7on in the source. were inadvertently pasted into the document.

Content is missing from the transla2on A paragraph present in the source is missing in the
12 Omission that is present in the source. transla2on

A source text states that a medicine should not be

administered in doses greater than 200 mg, but the
The target content does not accurately transla2on states that it should be administered in doses
13 Mistransla7on represent the source content. greater than 200 mg (i.e., nega2on has been omiMed).

The source text refers to a boy but is translated with a

Over- The target text is more specic than the word that applies only to young boys rather than the
14 transla7on source text more general term

The source text uses words that refer to a specic type of

Under- The target text is less specic than the miltary ocer but the target text refers to military
15 transla7on source text ocers in general

Content that should have been translated A sentence in a Japanese document translated into
16 Untranslated has been leR untranslated. English is leR in Japanese.

An transla2on is provided as an exact A TM system returns Press the Start buMon as an exact
Improper exact match from a transla2on memory (TM) (100%) match, when the proper transla2on should be
17 TM match system, but is actually incorrect. Press the Begin buMon.

Issues related to the form or content of a

text, irrespec2ve as to whether it is a A text has errors in it that prevent it from being
2 Fluency transla2on or not. understood.

Punctua2on is used incorrectly (for the An English text uses a semicolon where a comma should
21 Punctua7on locale or style) be used.

22 Spelling Issues related to spelling of words The German word Zustellung is spelled Zustetlugn.

Issues related to the grammar or syntax

of the text, other than spelling and
23 Grammar orthography. An English text reads The man was seeing the his wife.

The content uses the wrong gramma2cal

register, such as using informal pronouns
Gramma7cal or verb forms when their formal A text used for a highly formal announcement uses the
24 register counterparts are required. Norwegian du form instead of the expected De.

25 Inconsistency The text shows internal inconsistency. A text uses both app. and approx. for approximately.

21
 An HTML le contains numerous links to other HTML
les; some have been updated to reect the appropriate
Link/cross- language version while some point to the source
26 reference Links are inconsistent in the text language version.

Character
27 encoding

3 Terminology

Inconsistent A term is used inconsistently with a A termbase species that the term USB memory
31 with termbase specied termbase s7ck should be used, but the text uses USB ash drive.

The text refers to a component as the brake release

Inconsistent use Terminology is used in an inconsistent lever, brake disengagement lever, manual brake
32 of terminology manner within the text. release, and manual disengagement release.

4 Style

41 Awkward

The text violates company/organiza2on- Company style states that passive sentences may not be
42 Company style specic style guidelines. used but the text uses passive sentences.

Inconsistent One part of a text is wriMen in a light and terse style

43 style Style is inconsistent within a text while other sec2ons are wriMen in a more wordy style.

Third-party
44 style

45 Unidioma7c

There is a problem rela2ng to design

aspects (vs. linguis2c aspects) of the
5 Design content. A document is formaMed incorrectly

There is a signicant discrepancy

between the source and the target text An English sentence is 253 characters long but its German
51 Length lengths. transla2on is 51 characters long.

A por2on of the text displays a (non-systema2c)

Local Issues related to local formadng (rather formadng problem (e.g., a single heading is formaMed
52 formaSng than to overall layout concerns) incorrectly, even though other headings appear properly).

Issues related to markup (codes used

to represent structure or formadng of Markup is used incorrectly, resul2ng in incorrect
53 Markup text, also known as tags). formadng.

A transla2on is complete, but during DTP a text box was

Exis2ng text is missing in the nal laid- inadvertently moved o the page and so the translated
54 Missing text out version text does not appear in a rendered PDF version.

The German transla2on of an English string in a user

Trunca7on/ interface runs o the edge of a dialogue box and cannot
55 text expansion trunca2on-text-expansion be read.

A text document in UTF-8 encoding is opened as ISO

Locale Characters are garbled due to incorrect La2n-1, resul2ng in all upper ASCII characters being
6 conven7on applica2on of an encoding. garbled.

An online form translated from English to Hindi requires a

Content uses the wrong format for street number even though many addresses in India do
61 Address format addresses. not include a house number.

A text uses a date format inappropriate An English text has 2012-06-07 instead of the expected
62 Date format for its locale. 06/07/2012.

22
 A text dealing with business transac2ons from English
into Hindi assumes that all currencies will be expressed in
Currency Content uses the wrong format for simple units, while the conven2on in India is to give such
63 format currency. prices in lakh rupees (100,000 rupees)

Measurement A text uses a measurement format A text in France uses feet and inches and Fahrenheit
64 format inappropriate for its locale. temperatures.

A translated soRware product uses

shortcuts that do not conform to locale A soRware product uses CTRL-S to save a le in
expecta2ons or that make no sense for Hungarian, rather than the appropriate CTRL-M
65 Shortcut key the locale (for menteni).

A German text presents a telephone number in the

format (xxx) xxx - xxxx instead of the expected 0xx
Telephone Content uses the wrong form for followed by a group of digits separated into groups by
66 format telephone numbers spaces.

The text makes statements that The text states that a feature is present on a certain
7 Verity contradict the world of the text model of automobile when in fact it is not available.

An English text refers to steps in a process as First base,

Second base, and Third base, and to successful
Content inappropriately uses a culture- comple2on as a Home run and uses other metaphors
Culture-specic specic reference that will not be from baseball. These prove dicult to translate and
71 reference understandable to the intended audience confuse the target audience in Germany.

8 Other Any other issues

Severity levels for error categories

Severity 1 Cri2cal Errors that may carry health, safety, legal or nancial implica2ons, violate geopoli2cal
usage guidelines, damage the companys reputa2on, cause the applica2on to crash or nega2vely
modify/misrepresent the func2onality of a product or service, or which could be seen as
oensive.

Severity 2 Major Errors that may confuse or mislead the user or hinder proper use of the product/service
due to signicant change in meaning or because errors appear in a visible or important part of the
content.

Severity 3 Minor Errors that don't lead to loss of meaning and wouldn't confuse or mislead the user but
would be no2ced, would decrease stylis2c quality, uency or clarity, or would make the content
less appealing.

Severity 4 Neutral Used to log addi2onal informa2on, problems or changes to be made that dont count as
errors, e.g. they reect a reviewers choice or preferred style, they are repeated errors or
instruc2on/glossary changes not yet implemented, a change to be made that the translator is not
aware of.

Kudos Used to praise for excep2onal achievement.

23