For a bleeding-edge tech company, it’s sure hard not to feel Google’s been a bit slow to acknowledge and address a phenomenon that anyone in China has been witnessing for years — search results are censored.
In a new video and blog post, the folks at Google illustrate the problem and explain that they will now be notifying users that their search query may result in access to Google being cut off for a minute or more. The new feature also gives users options for changing characters (that have been shown to create temporary connection troubles to the search engine) to the less censor-offensive pinyin.
Two problems with this notice from Google though: the post is on Blogspot (Google’s blogging platform) and the video is on Youtube (Google’s video sharing platform) — neither of which are accessible in Mainland China without a VPN. So, to read the news you need a VPN, and if you have a VPN the news means little to you as you don’t suffer from that censorship. Bleeding-edge thinkers.
Slight dumbfounding aside, it is good information to know, and what they’re doing to address it is quite clever. For all of you with a VPN, you can go here and read the report. For the rest of you, here’s the whole thing (recopied without permission):
Over the past couple years, we’ve had a lot of feedback that Google Search from mainland China can be inconsistent and unreliable. It depends on the search query and browser, but users are regularly getting error messages like “This webpage is not available” or “The connection was reset.” And when that happens, people typically cannot use Google again for a minute or more. This video shows what’s happening:
We’ve taken a long, hard look at our systems and have not found any problems. However, after digging into user reports, we’ve noticed that these interruptions are closely correlated with searches for a particular subset of queries.
So starting today we’ll notify users in mainland China when they enter a keyword that may cause connection issues. By prompting people to revise their queries, we hope to reduce these disruptions and improve our user experience from mainland China. Of course, if users want to press ahead with their original queries they can carry on.
In order to figure out which keywords are causing problems, a team of engineers in the U.S. reviewed the 350,000 most popular search queries in China. In their research, they looked at multiple signals to identify the disruptive queries, and from there they identified specific terms at the root of the issue.
We’ve observed that many of the terms triggering error messages are simple everyday Chinese characters, which can have different meanings in different contexts. For example a search for the single character [江] (Jiāng, a common surname that also means “river”) causes a problem on its own, but 江 is also part of other common searches like [丽江] (Lijiang, the name of a city in Yunnan Province), [锦江之星] (the Jinjiang Star hotel chain), and [江苏移动] (Jiangsu Mobile, a mobile phone service). Likewise, searching for [周] (Zhōu, another common surname that also means “week”) triggers an error message, so including this character in other searches—like [周杰伦] (Jay Chou, the Taiwanese pop star), [周星驰] (Stephen Chow, a popular comedian from Hong Kong), or any publication that includes the word “week”—would also be problematic.
Now, when a user types in a common term like [长江] (Yangtze River) from China, Google highlights the problem term [江] as they type, and when they press “enter” a drop-down menu appears beneath the search box:
To learn more, users can click on the “interruption” link, which takes them to this help center article. They can continue with their original query (which will likely lead to an error message), or click “Edit search terms,” which will remove the highlighted characters and prompt users to try other search terms:
In order to avoid connection problems, users can refine their searches without the problem keywords. For example, instead of searching for [长江], they could search for [changjiang]—which also means Yangtze River, but is written using pinyin, the system used to transliterate Chinese characters into Latin script. This won’t cause a timeout, but will still generate search results related to the Yangtze River.
We’ve said before that we want as many people in the world as possible to have access to our services. Our hope is that these written notifications will help improve the search experience in mainland China. If you’re outside China and are curious to see what the notifications look like, you can visit this link to try it out.
Posted by Alan Eustace, Senior Vice President, Knowledge
Note: To read this blog post in Chinese, see this PDF.