Liesl L
Sr. Localization Strategist--Enabler of Multilingual Communications and Solutions
What software internationalization bugs have surprised you?
We're hosting another webinar this time on Software Internationalization (
http://www.translate.com/language_tech/webinars/Invitation_I18N_and_L10N_Webinar.html
) and I'd love to hear anyone else's experiences on internationalization bugs they've discovered.
I'll throw the first bug in. We had a client who is a prominent online dating site who rolled out a new platform. For some reason, the geography codes got mixed up and when people cancelled their memberships, the system would prompt them with "have you checked out these people from your area code". In Sweden, the system kept showing the men who had cancelled their memberships profiles of women from Thailand. Since there was a huge scandal several months before of international prostitution from Thailand, this lead to some embarassment.
Answers (4)
Paul D
Lead Software Development Engineer at RealNetworks
Best Answers in: Blogging (1), Software Development (1), Web Development (1)
Back when all of my projects were translated into 12 languages.
Most problems I've had were when character sets get mis-interpreted by browsers. Often breaking HTML or javascript.
My solution was to ensure all editors and translators produced content in UTF-8, then I wrote a program to convert all non 7 bit characters into entities. The entity always render the same, regardless of browser settings.
Other issues happened between designers and editorial. Designers would create mock-ups around English. Of course things broke when filled with German mega-sized words, or left with too much space when Japanese words were dropped in.
Robert D
interim management, Consultancy Microsoft Dynamics NAV, IT management,freelance, mind mapping, problem diagnosis,
Best Answers in: Enterprise Software (7), Organizational Development (6), Computers and Software (6), Business Development (2), Education and Schools (1), Mentoring (1), Corporate Taxes (1), Economics (1), Project Management (1), Small Business (1), Energy and Development (1)
Translation is incredibly difficult. Sometimes, languages get completely mixed up.
In our field we are largely confined to Windows platform, as most of specialized software is written for it.
The biggest problems come not even from work with any two-three European or Asian languages, but when the proper character encoding is to be ensured across more than two different platforms. Imagine working with Portuguese, Chinese, Arabic and Cyrillic scripts in one single document or web page.
Microsoft has been incredibly stubborn in insisting that Windows is an "international" platform. In fact, it is not, and simple choice of unicode or UTF-8 encoding does not solve the problem, especially when it comes to web programming. If, for instance, a website with some flash elements is to be translated, the programmer (or translator in charge of localization, for that matter) would virtually need as many different language versions of flash editor, as the number of languages of translation.
Same applies to database integration. It is not uncommon to use translation memory databases as part of dynamic web page or server-side includes. However, using any custom-made multilingual database, beyond rather primitive web translation tools and glossaries, so far has been an absolute nightmare.
This issue we probably cannot even call a bug, but a serious loophole.
To anyone who has done it - greeking can be a painful and laborious process to conduct.
A variety of issues can be found through greeking especially those areas of a product which prepare screen data or reports which do something like:
"Hello There " + strVariable + ", How are you today?".
Any product which will be internationalized should be built with composite strings which are prepared by adding some variable to the end of the string - not the middle. Without this - you may have a very difficult time translating.
-Robert
CTO/RevX Systems (www.revxsystems.com)
