Cultural Elements in Internet Software Localization

1
Cultural Elements in Internet Software
Localization

• Valentina Dagiene, Tatjana Jevsikova,
  dagiene_at_ktl.mii.lt tatjanaj_at_ktl.mii.lt
• Institute of Mathematics and Informatics
• Lithuania

2
Internet Software

• The term Internet software here is used as a
  general term to address
• software, used to access Internet resources
  (usually on a client side),
• web-based applications (server side).

3
Culture

• Provides the context in which the world is
  understood rules for behavior, communication,
  interaction and understanding.
• Multilevel onion-like models, e.g. basic
  assumptions and values, with resultant behavioral
  norms, attitudes and beliefs which manifest
  themselves in systems and institutions as well as
  behavioral patterns and non-behavioral items.
• There is a relation between users culture and
  software usability. Software can influence
  culture as well (this especially applies to
  Internet software).

4
Software Localization

• Software localization is software adaptation for
  particular cultural environment (locale).
• Unfortunately, still usually referred to as
  language translation.
• Localized software must look and feel as if it
  would have been made for the target language and
  culture.

5
Solving Culture-sensitive Issues

• At software production time, making it language-
  and culture-neutral and suitable for localization
  (internationalization).
• After software production time, modifying the
  original code at localization time.

6
An Aim of the Presentation

• To look at software elements that are based on
  culture and cultural conventions (a kind of
  reflection of cultural dimensions).
• To classify and discuss the most important
  software elements for successful cultural
  portability, basing on analysis of related
  normative documents and more than 10-year
  experience in software localization.

7
Classification

• A topic of studies by G. Hofstede, F. Trompenaar,
  E. Hall. The cultural dimensions identified by
  Hofstede offer possibility to structure culture
  according to the five concepts
• Power Distance.
• Individualism vs. Collectivism.
• Masculinity vs. Femininity.
• Uncertainty Avoidance.
• Long-term vs. Short-term Orientation.
• These are categories that organize general
  cultural data. Speaking about software, we can
  look at software elements that are based on
  culture and cultural conventions.

8
Structure of Cultural Elements in Software
9
Possible Users of the Classification

• Researchers to evaluate the level of
  internationalization of the original software,
  check the user-friendliness of localized
  software.
• Software developers to develop
  better-internationalized software.
• Localizers to adapt more cultural elements to
  the target culture and detect internationalization
  bugs.

10
Formal Definition of Cultural Elements

• International standard on procedures for
  registration of cultural elements (ISO/IEC 15897)
  defines locale as
• the definition of the subset of a users
  information technology environment that depends
  on language, territory, or other cultural
  customs.
• Locale is usually identified by the language,
  using two-letter language code (ISO 639-1), and
  by territory, using two-letter territory code
  (ISO 3166-1).

11
POSIX Locale Categories
12
Set of Formal Definitions of Cultural Conventions
(FDCC), ISO/IEC 14652

• format of postal addresses
• information on measurement system
• format of writing personal names
• format for telephone numbers and other telephone
  information.

13
International standard on procedures for
registration of cultural elements (ISO/IEC 15897)

• Specifies the procedures to be followed in
  preparing, publishing and maintaining a register
  of cultural specifications for computer use.
• First six clauses coincide with POSIX locale
  categories.
• Additional information national or cultural
  Information Technology terminology personal
  naming rules inflection hyphenation spelling
  numbering coding of national entities
  identification of persons and organizations
  electronic mail addresses keyboard layout
  man-machine dialogue, etc.

14
Unicode CLDR (more than 100 locales registered)

• Date and time formats
• Number and currency formats
• Measurement system
• Collation specification (sorting, searching,
  matching)
• Translated names for languages, territories,
  scripts, timezones, and currencies
• Script and characters used by a language.

15
Locale Implementation in Software
16
Locale Defined Elements(red rectangles)
17
Language-driven elementsAlphabets and Names

• Names (identifiers of various objects in Internet
  software, e.g. files, logins, passwords,
  domains...) are not only used by computers, but
  also by humans.
• Names in a native language and script are easier
  to
• devise,
• memorize,
• guess,
• understand,
• manipulate,
• correct, etc.

18
Restriction to use in names only English alphabet
letters (in outdated software)

• Forces a user not to use some/all letters from
  his/her native alphabet, but allow using foreign
  letters
• Most languages (even using Latin script), have
  some extra letters
• e.g., å, , ,
• Some English letters are not used in most of
  languages (using Latin script)
• usually q, w, and x.
• Makes impossible to use characters of non-Latin
  scripts.

19
The main reasons, why international characters
are not used in names today

• External some aspects of restriction for
  character use in names still exist in todays
  software.
• Internal previous experience on restriction had
  been applied for names affects users not to use
  national characters in names, unless such usage
  is technically possible.

20
Login Name

• Used in many web-based applications (virtual
  learning environments, e-mail clients, instant
  messengers, etc.).
• Characters
• Usually only underscores, numbers, and letters
  from the basic Latin alphabet are accepted.
• Some systems use the login name not only for
  internal identification but also for addressing
  the user in the system.

21
Personal Name

• Today, practically all the software allows using
  all letters of alphabet to write person's first
  and last name (surname) (a user shouldn't change
  or misspell his/her real name to register in the
  system).
• However, in telecommunications many users avoid
  using their native alphabet and write their names
  with spelling errors.
• For example, the number of incorrectly written
  names of Skype users varies from 10 to 90
  depending on the language.
• Such a great illiteracy may be caused by
  previous experience with outdated software or
  influence of present restriction on login names.

22
Passwords

• Used in software that performs users
  authorization (virtual learning environments,
  e-mail clients, instant messengers, etc.).
• Usually may be composed from letters and digits.
• Many programs still restrict the set of letters
  to ASCII alphabet.
• The restriction of the character set available
  for password reduces its security.

23
Passwords (an example)

• User usually does not think that letters in
  this context are only letters of English
  alphabet, but of his native language.

24
File/Folder Names for

• Storing documents on a local computer
• No technical problems in todays OS.
• Exchanging documents between computers by
  removable storage devices
• Works well as long as the same 8-bit encoding is
  used in both computers.
• Sending documents as parts of e-mail messages or
  as their attachments, or directly by instant
  messengers
• No technical problems. Before sending are encoded
  in UTF-8 (FF sequences) without non-ASCII
  letters, after receiving are decoded back.
• Storing web pages or other web content on a
  server
• Theoretically solved, the same method as sending
  by e-mail.
• Using inside applications
• A duty of developer to provide user-friendly
  names for visible items.

25
Domain names

• Till 2003 letters of Basic Latin alphabet (26
  letters), digits, dash.
• 2003 documents on using international characters
  in domain names were issued (RFC 3490, RFC 3491,
  RFC 3492)
• International characters (represented in Unicode)
  are converted to ASCII string (Punycode), and
  before showing it to user, it is converted back
  to Unicode characters again
• räksmörgås.josefsson.org ?
• xn--rksmrgs-5wao1o.josefsson.org
• Problems usage of homographs.

26
Domain names in browsers
27
Semantically-expressed elements Matching of
plural and singular forms

• English 2 forms
• 1 object, 2 objects, 10 objects
• Lithuanian, Polish, Russian ... 3 forms

• Some European languages, e.g. Slovenian, Maltese
  4 forms

28
Plural and Singular Forms in Other Languages
29
Grammatical Name Forms

• In inflective languages (Lithuanian, Finnish,
  Polish, etc.) names in dialog windows may appear
  in various cases.
• 'Hello, Jonas' (in English) will be
• 'Sveikas, Jonai' (in Lithuanian)

30
Gender

• S is logged in, S is a user name.
• English
• John is logged in.
• Mary is logged in.
• Lithuanian (and many other languages)
• John yra prisijunges.
• Mary yra prisijungusi.

31
Human-sensitive Elements

• Usually not defined by national or international
  standards (normative documents).
• Depend on deep cultural habits, country or its
  historical units cultural conventions.
• They can also depend on individual persons and
  should be adaptable to persons habits.
• They are difficult to express in a formal way
  (e.g. include into formal locale definition).

32
Some Examples

• Icons/Metaphors.
• Images, photos.
• Colour meaning.
• Usage of sounds and videos.
• Examples.
• Jokes and analogies.
• Political statements.
• Navigation scheme.
• Page layout.
• ...

33
Colour-Culture Chart(Boor Russo, 1993)
34
Icons Example Home Function
MS Internet Explorer
Possible Chinese icons
Mozilla Firefox
35
Problems

• Mentioned elements are more difficult to
  implement in internet software than in
  autonomously running software
• they are deeply grown into the program,
• internet software has many links with other
  software.
• Requirements
• flexibly adaptable to software and other cultural
  components
• flexibly fitting to each other
• flexibly chosen by the user (multiple choices).

36
Existing Ways of Solution

• Cultural Web Spider, designed to extract
  information on culture specific webpage design
  elements (cultural markers) from the HTML and CSS
  code of websites for a particular country domain,
  that could help to create a cultural interface
  design look and feel prototyping tool
  (Kondratova I., Goldfarb I., Gervais R.,
  Fournier, L., 2005).
• Many researchers confirm an importance of the
  cultural dimensions, set by Hofstede. They are
  used to create recommendations for a website
  navigation scheme and content presentation
  (Marcus A., Gould E.W., and others).

37
Existing Ways of Solution

• Recent research on incorporation cultural
  dimensions into global software includes attempts
  to create culturally adaptive software, applying
  AI mechanisms.
• It is also proposed to incorporate culture into a
  usermodel in order to implement adaptable
  personalization mechanisms, assigning Hofstedes
  value for each cultural dimension according to
  users birthplace, country of current and former
  residence, languages, sex, age, political
  orientation and education level (Reinecke K. et
  al, 2007).

38
Conclusions

• Existing shortcomings in software
  internationalization can be explained by the lack
  of categories included in formal locale
  definitions, and lack of compatibility of
  different locale models.
• While the developed list of cultural elements is
  limited, we hope that it can help to pay more
  attention to the complex set of cultural elements
  while designing, localizing and testing localized
  or intended to localize internet software.
• Special attention during internet software
  development should be paid not only for a
  generalized set of elements, defined in existing
  locale models, but also to the ability to use
  international characters in object names (names
  of logins, files, domains, passwords) an ability
  to include a component for languages grammatical
  forms generation usage of parameters in
  localizable strings should be reduced due to
  different rules of words and phrases composition
  in different cultures.
• Another trend for future work could be some
  formalization of human-sensitive elements, used
  in software.