Previous | Table of Contents | Next

3. TECHNOLOGY RESOURCES FOR MULTICULTURAL LIBRARY WEB SITES

This section of the toolkit presents information on the technology that can be used to develop multicultural Web sites. It also identifies and annotates critical Web resources on the technology of multicultural Web sites.

3.1 Multilingual Web Site Design

Designing multilingual Web sites is now becoming commonplace. A recent issue of The Economist ³ reported that of America's 100 largest firms, 57 had multilingual Web sites at the end of 2000, and that number is growing rapidly. Libraries have also started to design multilingual Web sites. Several examples of library Web sites with multicultural content are described in Section 5 of this toolkit.

One of the best Web sites for information on multilingual Web sites is LOTEit: Notes on a multilingual Web (http://www.yarranet.net.au/aceweb/LOTE/notesw95/notes.html). This is part of the Open Road site www.openroad.net.au, which was established by the State Library of Victoria in Australia and is based on the excellent work of Andrew Cunningham and Larry Stillman.

Since multilingual Web sites must deal with many languages it is important that the Web site be designed to easily accommodate them. There are several approaches to developing multicultural Web sites. These include:

• Providing a series of links on the library Web site to external sites in other languages. Instructions and directories on the library Web site are usually provided in either English or French. Using this approach, a library does not have to develop any specific, additional language or technical expertise. The difficulty of handling other languages and scripts is essentially downloaded to the external Web site.
• Providing information and brochures in other languages on the library Web site through the use of PDF files with embedded fonts or using multilingual keyboard support for common word processing systems. Again the actual content remains with the information providers on remote Web sites.
• Providing library Web sites that use Unicode (utf-8 encoded) Web pages designed to meet the needs of the multicultural communities in their own languages. These enhanced multicultural Web sites have to be able to handle different character sets and to provide quick translation systems. These sites provide multicultural information that is located on the library server. More information on Unicode and on machine translation systems is included further on in this section of the toolkit.

One of the basic requirements for any multicultural library service is to ensure that all multicultural materials are catalogued and are accessible through the library online catalogue. Most of the libraries described in Section 5 of this toolkit include materials in other languages in their online catalogues.

Andrew Cunningham provides a more detailed taxonomy of multicultural Web sites, which is based on access mechanism and degree of multilingualism. Cunningham points out that there are three types of access mechanisms for Web sites that provide multilingual resources.

• Resource level access. The Web site is organized according to the resources available. Most of the navigation is in a monolingual environment, usually English. All the translations into different community languages for that resource would be accessible from the same Web page. Resource level access facilitates Web site administration. All language versions of a document are located in the same directory. It also assists in updating and editing translations. This model is good for mediated access to information, i.e., when the information is being access on behalf of someone.
• Language level access. The Web site is organized by language. This structure looks logical to the person browsing the site, but increases the amount of administration required. In a poorly designed site, it may lead to duplication of files and images throughout the site. It would be difficult to manage as a large site. Language level access is a preferable access mechanism for people with low English literacy skills.
• Hybrid access model. The site structure is based on the resource level access model, allowing for streamlined site management coupled with language site indexes or search mechanism so that the person using the site can search in their own language for resources in their language.

Cunningham also divides multilingual Web sites into three categories:

• "Monolingual" (Unilingual): the site is multilingual, and each page is unilingual. Unilingual Web pages can be of any length.
• Bilingual: each page contains two languages. Bilingual pages usually contain only one or two paragraphs in each language.
• n-lingual: each Web page contains multiple languages. n-lingual pages tend to contain short sections of text, no more than one paragraph per language. If each language had more than one paragraph of text, the page would become difficult to update. It is common to use an n-lingual page as a navigational tool in a frame.

There are a growing number of Web sites that present information and services from both commercial and non-profit organizations for developing multicultural Web sites. Some examples that demonstrate the types of information and services that are available follow:

Babel
http://alis.isoc.org/index.en.html
Babel is a joint initiative of Alis Technologies and the Internet Society and is aimed at internationalizing the Internet. The site is presented in English, French, German, Italian, Portuguese, Spanish and Swedish. Babel includes information on the major languages of the world from an Internet perspective, and a technical section dealing with Internet internationalization, including document coding in various languages, and information on developing a multilingual Web site.

Multilingual Application Interface for Telematic Services (MAITS)
http://std.dkuug.dk/maits/
MAITS is a consortium formed to specify an Applications Programming Interface (API) for multilingual applications in the telematic services. MAITS aims to integrate and extend the various solutions for dealing with locales and character sets that already exist, notably from the X/Open consortium. MAITS plans to develop an Application Programming Interface (API), covering four language processing levels:

• character set conversion between client and server;
• transliteration and locales;
• simple translation of stored text strings; and
• access to machine translation.

MultilingualWebmaster.com
www.multilingualwebmaster.com
MultilingualWebmaster.com provides an open forum for developers and managers of multilingual Web sites. The site provides information on best practices in multilingual Web site design.

SDL International - SDLWorkFlow, Multilingual Content Management System
www.sdlintl.com/sdlworkflow
SDLWorkFlow provides support for automating the work relating to creating and managing multilingual content on Web sites by automatically routing new content or changes through multi-level workflows and applying "Translation Memory" and "Real-time Translation" software.

World Wide Web Consortium (W3C)
www.w3.org/International/
The goal of the W3C Internationalization Activity is to propose and coordinate techniques, conventions, guidelines and activities within the W3C and together with other organizations that allow and make it easy to use W3C technology worldwide, with different languages, scripts and cultures. The Web site provides significant Web internationalization resources including:

• information on creating Web pages with non-western character sets, languages, and writing systems; and
• list of languages and the character sets commonly used.

Unicode^®
www.unicode.org/
Unicode is a multilingual and multiscript character set, i.e., a coding system that allows the representation on screen of over 65 000 characters. The Unicode Standard defines codes for characters used in all the major languages written today. Scripts include the European alphabetic scripts, Middle Eastern right-to-left scripts, and many Asian scripts. The use of Unicode makes it possible to create multilingual Web pages in many different languages. Library and Archives Canada plans to use Unicode to provide access in the language of the publications in their collections, especially those in non-Roman scripts.

The Web site is managed by the Unicode Consortium, a non-profit organization founded to develop, extend and promote use of the Unicode Standard, which specifies the representation of text in modern software products and standards.

For an introduction to Unicode see:
www.unicode.org/unicode/standard/principles.html.

For an excellent introduction on the use of Unicode in developing multilingual library Web sites see Andrew Cunningham's article "Multilingual Unicode Web Page Development" located at http://members.ozemail.com.au/~andjc/papers/cn99.html.

3.2 Multilingual and International Search Engines

All the major search engines have developed or are developing multilingual capacity, thus allowing the user to search the Web in many different languages. The multilingual search engines and the multilingual e-content available on the Web make it feasible to build multilingual portals in major community languages enabling library users to find and access these resources.

This section provides a list of Web sites that provide an introduction and access to multilingual and international Web sites.

Toronto Public Library - International Search Engines and Directories
www.tpl.toronto.on.ca/mul_bks_search_engines.jsp
The Toronto Public Library Web site provides access to an extensive list of search engines under the following headings:

Searching by Topic

• Subject Directories
• Subject Directories - Canadian Focus
• The Invisible Web - Databases and Gateway Sites

Searching by Keyword

• Search Engines
• Meta Search Engines

French Language Directories & Search Engines

International Directories & Search Engines
http://vrl.tpl.toronto.on.ca/internet/01_engin_1_c.html
#International%20Directories

This page links with "Country-Based Search Engines"
Country-Based Search Engines
www.philb.com/countryse.htm

This site provides access to country search engines and regional search engines - currently a total of 1 533 search engines and 169 countries and regions.

Vancouver Public Library - International Search Engines
www.vpl.vancouver.bc.ca/branches/LibrarySquare/multi/links.html/
The Vancouver Public Library site provides links to

• Global Excite, available in Italian and Japanese;
• Go International, available in Japanese;
• Lycos, available in Chinese, Danish, Dutch (Netherlands), French, German, Italian, Japanese, Korean, Spanish and Swedish; and
• World Yahoo, available in Chinese, Danish, French, German, Italian, Japanese, Korean, Norwegian, Spanish and Swedish.

Google Language Tools
www.google.com/language_tools
The "Google in Your Language" site provides access to

• Search pages written in any language (74 languages); and
• Search pages located in any country.

Searchengine.Colossus
www.searchenginecolossus.com
A gateway to country-specific directories of links. It offers links to search engines from 172 countries around the world.

Search Engines World Wide
http://home.inter.net/takakuwa/search/search.html
Search Engines World Wide provides a list of global and local search engines sorted by country. As of January 7, 2003, the site provides links with search engines worldwide as follows:

Yahoo
www.yahoo.com
Yahoo provides 24 country or language specific search options.

Terra Lycos World Wide
www.lycos.com
www.terralycos.com/
Lycos provides 42 country or language specific search options.

With the ever-increasing availability of multilingual search engines, there no need for libraries to index and create Web directories for non-English Web site for most languages. Web directories may still be required only for small, minority languages that have a limited Web presence and are currently unsupported by existing Web directories and search engines.

3.3 Translation Technology

Globalization has dramatically increased the demand for translation services for governments, businesses and individuals. There are not enough human translators to meet the demand economically. This has generated considerable interest in developing machine translation software. Although machine translation is still far from perfect, it has opened the door to actual multilingual conversations. Libraries that are engaged in developing multicultural Web portals should include translation software on their sites. This is the only way that information available in one language can be made available to someone who speaks another language. Translation technology is essential for a truly multicultural service, as it is the only way to handle large amounts of multilingual information.

Accurate machine translation systems depend on having access to translation assets, that is, past translations, translation memory, style guides and terminology databases. In order for machine translation systems located on multilingual library portals to be effective, libraries will have to develop their own translation memory. Any single library is unlikely to be able to develop a critical mass of library-related translation memory. This can only be done through libraries working cooperatively on multilingual Web portals. Therefore, although the short-term goal may be for libraries to develop their own multilingual Web portals, the long-term goal must be collaborative multilingual Web site development.

The following is a brief description of some of the many available translation technology solutions.

AltaVista - Babel Fish Translation
http://babelfish.altavista.com/
The AltaVista search engine provides a free translation service based on Babel Fish software from Systran. Translation is provided between English and nine major languages.

Bowne Global Solutions
www.bowneglobal.com
Bowne specializes in supporting corporate globalization plans and provides multilingual end-to-end publishing processes tailored for any culture and language. Bowne also provides a plug-in for Word 2002 that provides instant translations of Word content without leaving the application. Using translation software, a user can elect to have a document translated into a variety of languages from the pull down menu at no charge.

Dictionary.com
http://dictionary.reference.com/translate/text.html
Dictionary.com provides a free translation service in English, French, German, Italian, Norwegian, Portuguese and Spanish, based on Systran technologies.

Foreign Word
www.foreignword.com/
Free text translation from more than 30 languages, including Afrikaans, Chinese, Italian, Swahili, Malay and Ukrainian, into English.

FreeTranslation
www.freetranslation.com/
FreeTranslation.com is an easy-to-use site for rapid translations, where you can get the "gist" of foreign language text and Web pages in English, French, German, Italian, Norwegian, Portuguese and Spanish.

Google
http://translate.google.com/translate_t
Free "gist" translation service for English, French, German, Italian, Portuguese and Spanish. "Gist" is used to describe partial translations, which enable a person to get the 'gist' of the item being translated. Google uses Systran technology.

Langenburg.com
http://translation.langenberg.com/
Provides links to language translation services in English, French, German, Danish, Spanish, Finnish, Czech and Italian.

The Multilingual Translation Technology Centre, University of Toronto Libraries
www.library.utoronto.ca/translation
The centre consists of two PCs, a laser printer, and a professional flatbed scanner (HP Scanjet 6300) with an automatic document feeder for multiple-paged documents. The following text translation software is available:

• Universal Translator is a simple computer translation program that lets you translate text bi-directionally in 40 languages.
• Systran Enterprise 2.0 is a beta version of the most recent and most powerful computer-assisted translation program. Translates text in English, French, German, Italian, Japanese, Portuguese, Chinese and Russian. Also includes 21 specialized glossaries. www.systransoft.com/Corporate/Enterprise.html
• Systran Professional Premium is machine translation for bi-directional English, French, German, Italian, Japanese, Portuguese, Russian and Chinese into English. (No English into Chinese or Russian.) Also includes 20 specialized glossaries.

Systran Information and Translation Technologies
www.systransoft.com/
Systran provides scaleable and modular translation architecture. Its core technology delivers translation solutions for the Internet, PCs and network infrastructures that facilitate communication in 35 language pairs and in 20 specialized domains. Systran technology is behind the translation services offered by the following portals: Google, AOL, AltaVista, AOL, Apple's Sherlock Internet Search, CompuServe, Lycos and OracleMobile.com. The Systran Web site provides a free quick translation service in 10 languages, including Chinese, Japanese and Korean for text and for Web sites.

Translation Experts
www.tranexp.com/
Translation Experts provides translation software and services for translating Web pages and e-mail messages in over 30 languages. InterTran, their Web site translation server runs on Windows XP, 2000, NT, Solaris and Linux and will translate Web pages in Albanian, Bosnian, Brazilian Portuguese, Bulgarian, Chinese, Croatian, Czech, Danish, Dutch, English, European Portuguese, Filipino/ Philipino/Tagalog, Finnish, Flemish, French, German, Greek, Hindi (transliteration), Hungarian, Icelandic, Italian, Japanese, Korean, Norwegian, Polish, Rumanian/Romanian, Russian, Serbian, Slovenian, Spanish, Swedish, Turkish, Ukrainian, Vietnamese and Welsh.

Transparent.com - Easy Translator
www.transparent.com/
Easy Translator provides quick translation for documents, text files, e-mail and documents and Web pages in English, Spanish, French, German, Portuguese and Japanese.

World Language
www.worldlanguage.com/Translation.htm
Free automated translations that are not perfect, but they can give you a general idea of what is being communicated. Some of the languages have more limited lexicons than others, so the quality of translation can vary between the languages. Some of the languages will need additional support for specific fonts. The site provides text and Web page translation between English and approximately 30 European languages.

WorldLingo
www.worldlingo.com/
WorldLingo offers services in a range of languages and provides both human and machine-based translation services. WorldLingo provides e-mail translation systems and Web site translation systems.

Their Instant Web site Translator will translate any English Web site into French, German, Italian, Spanish, Portuguese, Dutch, Greek, Korean, Chinese and Japanese.

WorldLingo's Chat capability enables interactive, real-time multilingual discussions. A user writes a message in his or her language, and their chat partner will instantly see that message in both the user's language and theirs. Users can choose to view the message in multiple languages. Multilingual Chat offers translation to and from a range of language pairs including English, French, German, Italian, Spanish, Portuguese, Japanese, Korean, Chinese and Russian.

Microsoft has built in WorldLingo's translation software into Office XP and Word 2002. Microsoft's translation tool provides translation services for English, French and Spanish and automatic downloads of translation software from WorldLingo and Bowne Global Systems.

WorldLingo is one of the companies working on the development of translation memory systems. Their system is known as Translation Asset Management System.

Some other translation systems include:
Softissimo - Reverso Pro 5 www.softissimo.com/
IBM - Websphere Translation Server 2.0
Terminotix - LogiTerm www.terminotix.com/
WebViews - Inter Language Paradigm www.web-views.com/

³ "Tongues of the web," The Economist, May 14, 2002.

Previous | Table of Contents | Next