Web Globalization On A Local Budget
Published on December 3, 2003
When we think of user-centered design we have yet another user to consider in our designs—the rest of the world. We have all seen large Web sites such as BBCs World Service, IBM, Macromedia and others that have not only created sites in other languages (internationalization), but also customized their content to the particular region they are targeting (localization). Essentially they have created multiple versions of their site, which requires huge financial resources and some serious content management. But what about the rest of us who don’t have the financial resources or time to maintain multiple versions of our sites? Although it would be ideal to have our sites translated into 43 languages, for most of us that don’t have a defense budget (if any budget at all) it is impossible. However there are a number of free tools available to you that, in combination with a little bit of professional translation and the utilization of an XHTML/CSS-based design, will make your site accessible to the rest of the world.
The cheapest and easiest way to translate content is through the use of machine translators and there are quite a few services that allow you to use this service for free such as Google, Alta Vista’s Babelfish, Alis and World Lingo. My personal preference is World Lingo, primarily due to their ease of use, language choices, and the fact that you can select the subject matter of your content: as a result, the translation will be a bit more accurate. So let’s say I decide to use World Lingo and I want to translate my blog into Dutch, first I run it through their machine translation service and then I copy and paste the URL (may need to view source) to my link and now I have a link to my blog in Dutch.
Web Standards Help the World Sing in Harmony
I think the benefits of using Web standards has been pretty much beaten into our heads. We have heard about the benefits such as greater accessibility for those with disabilities, smaller file sizes, cleaner code and a move towards the separation of content and design. Well, another benefit of using an XHTML/CSS-based design is that it makes it easier to implement multi-lingual content.
One of the biggest headaches in dealing with multi-lingual content used to be dealing with different character sets for each language, for example a few years ago if I wanted to have a page in Japanese, Korean and English I would need to use a different character set for each page:
<meta http-equiv="Content-Type" content="text/html; charset=Shift_JIS" />
<meta http-equiv="Content-Type" content="text/html; charset=euc-kr" />
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" />
Luckily with current browsers we no longer need to go this route, as the new default character set for the Web is Unicode. The standard ASCII character set only contains 256 characters, however, Unicode has 94,140 characters which include characters for all of the major languages including Chinese, Russian, Japanese, Korean, Arabic and the list goes on. Although Unicode itself is just one character set, it has multiple encodings; the standard encoding is UTF-8 (UTF stands for Unicode Transformational Format), as it is, for the most part, backwards-compatible with most servers and routers that can only handle single byte characters. So if you put the following in the header of your HTML document, you will have all those languages covered:
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
This makes things a great deal easier from a development standpoint, especially if you are working with dynamic pages generated by the server. We also no longer need to do any language detection, or worse, hacks such as asking the user to switch character sets in their browsers.
In order for you to develop with Unicode and multiple languages you will need to have a system and software that can handle it. For my setup I use Windows XP with the East Asian Language pack installed which allows me not only to read Asian characters but also to write them. You will also need a text editor or IDE that supports Unicode, I use DreamWeaver MX which handles Unicode and multi-lingual content perfectly, however there are quite a few free Unicode text editors available out there, as well. If you are using Windows XP you can enable Asian languages by going to the Regional and Language options in your control panel, clicking the languages tab and click the checkbox next to “Install files for East Asian Languages”. Click OK and it will prompt you for your Windows CD to install the language packs. There are a number of software pacakages available, such as one from Twin Bridge that can be purchased and installed on older systems, or if you need to be able to use these characters in software that does not work well with Unicode.
Another benefit of using an XHTML/CSS-based layout is the ability to move away from graphics for your navigation. If you plan on machine-translating content this will allow you not only to translate your content but also your navigation. For example if I take a page that uses CSS-styled navigation and run it through a machine translator into Japanese the navigation will be translated as well as the content. Although the translation is a bit off, and in some cases will be wrong, for those users who are not fluent in English it is better than nothing and as they click each link the machine translator will automatically translate each page as they go through. If you have the appropriate language packs installed on your machine you will see something like this:
Don’t Trust Machines
Although the use of machine translations is a step in the right direction, it is not always the best solution to use by itself. The main problem with machine translations is the fact that you are trusting your content to a program. When a page is translated from one language to another some things will not translate as well as others. Another thing to consider is that metaphors, phrases and humor are generally culturally specific and when translated into another language may sound a bit silly, or could even potentially be offensive. This may not be a big deal if it is a personal site, but if it is a business site where you are trying to sell a product or service you may want to think twice about relying solely on the use of machine translations.
So if we can’t afford to translate our entire site and can’t rely solely on machine translations, what’s a developer to do? One thing that can be done is to simply have one page professionally translated with the core information that an international visitor will need to know about your site, business, products and services. On this page you can then also explain to the user that the rest of the site is in English and apologize for the inconvenience, but if they click a link that you provide it will take them to a machine translated version of your site. The cost will generally be around $150-$200 for each language/page depending on the language and length if you go through a professional translation service such as World Lingo or Alis. You may even be able to get it done cheaper by posting an ad at a local university and working with either a foreign student, or a student majoring in the language you want to have your page translated into. I have been pretty lucky in that I have had good results working with professors at the university I work for and also my wife who is trilingual. I would simply send the English version of the pages off in Word to the translator and they would send the translated version back to me in Word. I would then create a Web page using UTF-8 as my character set, copy and paste in the content, cleaning it up a bit along the way until the final product looks something like this.
Since I used Unicode you can see the content in English, and also the Japanese and Korean links together on the bottom left, all of which are simply CSS-styled links. The box on the right-hand side simply tells the user that we apologize for the rest of the site being in English but that if they click the link it will give them a machine-translated version of our site in their language.
In order to get our foreign visitors to this page I created a graphic and placed it on the Home page. The reason for using a graphic, as opposed to text, was aesthetic. Since the majority of our users do not have Asian characters installed on their system, the result could look pretty ugly. I also kept the link low-contrast in relation to the rest of the site since having the text in the user’s native language is enough to draw the eye to it. Imagine going to a site that is entirely in Chinese, but if, in the upper right-hand corner, you see a small link that says “English”, where to click becomes quite evident. I also put the graphical language link in our global footer, so on every page there is a link to the Korean and Japanese pages.
The main reason that we currently only have Japanese and Korean is primarily due to the fact that after three months of logging our users’ default languages, Japanese and Korean were the top languages aside from English. The entire project cost me $0 as I was lucky enough to have people volunteer their time to assist with the translation. However, we plan to expand these pages to other languages in the very near future to include Chinese, Spanish, French, Portuguese and Arabic. Even if you pay to have one page translated into one or more languages, the cost will be small in relation to the possible benefits. For Pacific University, the majority of foreign visitors will probably know enough English to get through the site without the additional assistance. However, the fact that we take the time to consider them, and make their visit easier, will make those students feel more welcome at our institution and help those who may not be as fluent to get to the information they need about our English Language Institute. Adding links in multiple languages to your site may also make your company or institution look more distinguished and international in comparison to your competition.