Google fixes China search bugs

An investigation published Thursday by CNET News.com showed that Google's new China search engine not only censored criticisms of the Chinese government, but went further than similar services from Microsoft and Yahoo by targeting sites related to teen pregnancy, alcohol, dating and homosexuality.

On Friday morning, however, those previously verboten sites became available through Google.cn. That brings Google's filtering in line with blacklists used by Microsoft and Yahoo.

Google did not immediately respond to a request for comment on Friday. Earlier in the week, however, the company had said that its product launches are typically "followed by a process of identifying and correcting bugs or other technical issues. Google.cn is no exception, and we will continue to refine our processes to ensure that we are filtering the minimum necessary."

"Before this, I was actually of the mind that if you're doing the disclosure, this is the way to go," said Danny Sullivan, editor of SearchEngineWatch.com. "The more I thought about this, I realized this is evil. You're caving into censorship that you in your heart of hearts don't believe in."

Even though Google founder Sergey Brin had promised that deletions would be listed, the initial version of Google.cn frequently did not do so. The debugged version appears to have fixed that problem.

It's not entirely clear what caused the initial version of Google.cn's embarrassing bugs, which had caused Microsoft sites and the University of Pennsylvania's engineering school to vanish from search listings.

One potential explanation is that Google had been using a version of its "SafeSearch" software, which it describes as adult-site filtering technology that "checks keywords and phrases."

SafeSearch has proven to be flawed in the past. A 2004 investigation by CNET News.com showed that it blocked many innocuous Web sites based solely on strings of letters such as "sex," "girls" or "porn" embedded in their domain names. PartsExpress.com, ALittleGirlsBoutique.com, RomansInSussex.co.uk, ArkansasExtermination.com and BassExpert.com, for instance, were incorrectly identified as pornographic.

Google.cn seemed to work the same way, delisting an Essex County Web site apparently because the word "sex" appeared in the the domain name.

Because access from China to the U.S.-based Google.com site is limited for financial and political reasons, the vast majority of Chinese are forced to turn to domestic search engines instead. Google's Brin has estimated that Google.com is available to only half of the country's users. Other reports say that when search terms such as "Tiananmen Square" are typed in on Google.com, the site immediately becomes unreachable for a few hours.