Internet Infrastructure and the Lessons Learnt from Web Localization

1997 was marked by the continued growth of the Web and the progressive spread of the Internet infrastructure. The Asia Pacific Economic Review reports [1] that "as of 1996 the epicenter for Internet infrastructure growth began shifting strongly toward the Asia Pacific Region", quoting the two fastest growth countries to be Hong Kong (now part of China) and Japan, with growth rates of 178% and 173% respectively. At the current rate of growth "by the end of the decade, Asia will have 25 million to 30 million users on-line". While the recent downturn in some Asian economies may dampen the rate of growth in the short term, in the long run the communications infrastructure will relentlessly progress in the direction of the global information society supported by computers and telecommunications.

The Internet infrastructure has introduced a range of new modalities of communication such as e-mail, online chat and the Web, all of which are having a widespread impact on society. The Web is now used by the entertainment industry (almost all new Hollywood films have their own Websites), commerce (e.g. telebanking, online shopping and teletranslation) and edu-cational institutions (e.g. the online library, virtual university, etc.) The fact that the Web is instantaneously global seems so obvious and yet it took some time before this significant implication sank into corporate thinking: a homepage provided in a single language may be blocking communication with potential customers accessing from diverse linguistic and cultural backgrounds. Richard Tuttle [2], Director of the International Webmaster Association Language Center, argues that "Some day soon, we believe every movie's Website will be translated into five or more languages. Why not? The movies themselves are." The rapid increase in Web localization efforts during 1997 is a direct consequence of the realization of the Web's global reach. Today not only multinationals but also much smaller businesses have launched their Websites in different languages. Now that the new infrastructure is through its embryonic stage, we are beginning to grasp the nature of the new environment.

This has meant a steep learning curve for localizers and translators. Of particular relevance is Asian localization of Websites which demanded overcoming the dual compatibility barriers of double-byte characters and elusive cultural issues [3]. This has taken the localizer's role from transcoding to that of boutique translation, in which not only the language code is switched, but its information content and overall presentation, such as design and layout, are also taken into consideration. Maturing Web localization, both from the users' and the suppliers' points of view, is evident in some multilingual sites. For example, the Asia Pacific Economic Review singled out [4] Citibank's multilingual sites to be "sensitive to the local audience" as they present "different faces and customized imagery and text for each site". Its Japanese site (www.citibank.com/japan/ gcb/ja/index.htm) certainly displays a number of "typical Japanese" formulas such as the use of the cartoon, choice of color, etc. This contrasts with its Singapore site where photographic images of male and female "guides" appear (most European and Australian sites have images of female bank tellers alone). Similarly, Visa's international sites indicate a market-tailored approach. For example, its Japanese site (www.visa.co.jp) highlights the forthcoming Nagano Winter Olympics, while the German site has a link to the latest James Bond film (presumably because the film features Germany and some German characters).

Microsoft also takes the "local reader"-oriented strategy for its Web localization, as reported in a recent issue of MultiLingual Communications and Technology magazine [5]. According to the article, Microsoft's overseas subsidiaries, working with their existing local localization vendors, take responsibility for supplying local content for the relevant Website while product-specific localization is primarily carried out centrally by their in-house localization team in Dublin. By comparison with the Citibank sites, however, Microsoft's sites come across as fairly uniform to its American headquarters' site in style and presentation. This is explained by the fact that their Web material is "largely product- rather than market-specific". Different products beg different Web strategies, but the common approach in these multilingual sites is to maximize the value added by the Web as a global medium by making the content and presentation relevant to the target reader.

The Wordless Web and the Emergence of VRML

So, what is the next phase of the Web's development? Today's Web is still largely a text-based medium as the name of its core mechanism, "hypertext", illustrates; it remains a text-centric, flat world. However, Michael Schrage [6] suggests that "language is going to play less and less of a role as the Web becomes more and more global. We will slowly see a shift from language-centric to iconic communication. Just as we have global traffic signs, we'll have global Internet icons." As he points out, the precedent for this is the use of GUIs (Graphical User Interfaces) in the computing world. Schrage is suggesting that "pictures" as a sort of Web Esperanto will eliminate the need for words. The use of cartoon characters and avatars may also be contributing to what Schrage calls "the Wordless Web" that "relies on clever icons and representations rather than wordy lists of choices".

His comments on the increasing importance of iconic communication are significant in relation to the next wave in the Web evolution. One emerging direction on the Web is the use of three dimensional (3-D) interactive graphics based on VRML (Virtual Reality Modeling Language), a standard for describing 3-D space, allowing interactive 3-D graphics to be embedded in Web environments. The VRML-based tools enable the creation of such environments as 3-D "information landscapes" which visitors can "walk through" (virtually), with perspectives changing according to their relative positions, and where they are able to build their own 3-D objects. Since early 1997, numerous 3-D browser plug-ins have been distributed to enable Net surfers to view 3-D scenes in real time. VRML seeks to be part of the Internet infrastructure to establish the Internet VR with such capabilities as dynamic information visualization and distributed collaborative environments on the Internet.

Applications of this new environment span from entertainment to medicine, engineering and education, as well as commerce. The shift from HTML to VRML signifies a move to a 3-D land of spatial communication. Its high-end applications will enable, for example, industrial product designers in various locations to meet in a 3-D virtual world and discuss and modify in real time product models presented in interactive 3-D graphics, while students at the virtual medical school will practice for surgery on a 3-D model of the human brain. With progressive improvements on the information superhighway and refinements of VRML, the speed of interactions and accuracy of the graphics will increase. VRML with its newly found means for spatial communication will change the way we interact with information and people online, perhaps making us less reliant on words.

However, on another level this new capability will mean that cultural differences could become more visible as the Internet VR develops to accommodate such elements as non-verbal communication. For example, today's effective, albeit primitive, "emoticons" (e.g. ;-)) could be replaced by more sophisticated kinetics using avatars, providing context normally expressed by extralinguistic means such as facial expressions, gestures, or seating positions. In a virtual world one will have control over the presentation of these and other far more subtle forms of expression: dress code, direction of gaze, sign of appreciation, etc. Although there is some universal body language, localizers and translators need to be acutely aware of cultural differences in these non-verbal communication cues. For example, Americans normally keep at arm's length while standing and talking, but in South America, the Mediterranean region and the Middle East people sometimes literally stand toe-to-toe [7]. These behaviors may become important parts of communicating on the Internet.

Implications for Localization and Translation: A Chinese Bill Gates?

As touched on earlier, today's Web localization is mainly text-based, with a few additional non-textual features. With VRML is this balance going to flip? Certainly today's Web users all know that the online environment is not the best for reading any lengthy or complicated document; we only "browse" the document online and print it out to digest offline. This practice will continue until the basic mismatch of text and today's computer screen can be overcome. Meanwhile, growth in the interactive online content of the Internet will come through increasing use of images, sounds and, further into the future, tactile senses. The reference to avatar developments seems relevant at this point.

Early avatars described by Rheingold [8] in Habitat were cartoon-like figures with movable heads and bodies. In the emerging Internet VR, these avatars are increasing in sophistication, and have a growing range of facial expressions and gestures. For example, a visit to Internet VR sites such as AlphaWorld (www.activeworlds.com) can give visitors the feel for today's state-of-the-art avatars. Furthermore, the virtual environments where these avatars live are also becoming elaborate. Using even today's technology the still images (photographs and cartoons) of tellers at a bank's Website could be replaced with a smiling 3-D avatar who greets and directs visitors with corresponding gestures. In the near future avatars and virtual environments will be localized/translated in culturally compatible forms which include appearance, gestures and any other relevant behaviors. A Japanese female bank teller will have certain expected mannerisms which are different from those of her counterpart in Australia. In some cases the differences in non-verbal communication from one culture to another can be quite drastic.

This consideration is also relevant to the design of icons in today's computing environment, and particularly for localization. Web localization with these virtual environments will become an increasingly complex exercise, highlighting new kinds of cultural considerations beyond words. In future, avatars representing Bill Gates will come out behaving differently as a Chinese avatar from his original American form and different again as a German, and the differences must be more than cosmetic. This is the kind of subtle but deliberate manipulation that human interpreters habitually perform; they may spice up a British speaker's presentation for a Spanish audience, while toning down an extremely expressive and animated speech given by a Spaniard to a British audience. While more use of global icons in the Web may in some cases remove the need for localization of text, the use of images and the move to spatial communication will create an altogether different type of challenge.

A Vision: What Will Be?

The title of this article was inspired by MIT Computer Science Lab Director Professor Dertouzos' latest book What Will Be [9], in which he discusses predicted developments in the coming information revolution based on his technical expertise in computing and key MIT research projects. He argues that "the Information Marketplace" will facilitate our wide-ranging activities in work, health care, entertainment and education, allowing interactions at a distance on a global scale. He maintains that the prevalence of groupwork and telework modules, linking people from diverse locations, will grow and the benefits of "electronic proximity" for the elderly will be great. In future, people may even be able to send their "experience" (as opposed to mere text) in an e-mail message, which the recipient in turn will be able to "feel". In this way, he maintains, the Information Marketplace will superimpose a "cultural veneer" of shared experiences on top of the individual cultures of the world. My point is that this very creation of the shared experiences among people with no common language and cultural backgrounds necessitates localizers' and translators' contributions.

As today's text-centric Web moves to embedded VR, our electronic interactions will involve increasing amounts of synchronous interlingual and cross-cultural communication. So far, localization has catered well to intralingual communication by providing pre-translated environments for people who stay within the same language paradigm. The future task for localization, however, may be to deal with high-speed hyperlinks from one cultural landscape to another. How should a multiparty multilingual VR chat session be most effectively mediated? Will localized products need to embed some basic MT systems? Would such systems be adequate in all circumstances? When and how should on-line human language assistance be invoked? How could teletranslation be used? In my current research I am trying to answer some of these questions. Our job as localizers and translators is to make the Information Marketplace accessible, unhindered by linguistic and cultural barriers.

References

[1] "Localizing to Globalize the Internet" Asia Pacific Economic Review (1997, Summer, Vol. V, No. 2)

[2] Tuttle, R. [1997] "WebSite Translation: An introduction for Webmasters, HTML authors and WebSite Owners" from www.irwanet.org/languagecentre/

[3] See my article "Rising to the Challenge of Asian Localization and Translation" in LISA Newsletter (August 1997, Vol. VI, No. 3)

[4] Asia Pacific Economic Review (Ibid.)

[5] Bussin, R. [1998] "Microsoft's Multilingual Web Site Strategy" MultiLingual Communications and Technology (Vol. 9 No. 1)

[6] Schrage, M. [1996] "The Wordless Web" from www.hotwired.com/packet/ schrage/96/46/index1a.html (Note, he is the author of "Shared Minds: New Technologies of Collaboration")

[7] Axtell, R. [1991] "Gestures" New York: John Wiley & Sons, Inc.

[8] Rheingold, H. [1995] "The Virtual Community" London: Minerva.

[9] Dertouzos, M. [1997] "What Will Be" New York: HaperEdge.

---

Minako O’Hagan

School of Communications
and Information Management,
Victoria University of Wellington,
New Zealand

E-mail: Minako.OHagan@vuw.ac.nz