Hacker News new | past | comments | ask | show | jobs | submit login
Google will start warning Chinese users about problematic search terms (insidesearch.blogspot.com)
132 points by raldi on May 31, 2012 | hide | past | favorite | 58 comments



This is totally badass. Instead of a technical solution to a technical problem, Google is going with a political solution to a political problem.

There are a ton of technical ways Google could silently defend against this particular form of censorship (someone suggested reencoding the offending characters), thus starting a cat-and-mouse game with the censors. But that would (1) be an "act of war" in terms of blatantly attempting to pierce China's Great Firewall, giving China an incentive just to block Google wholesale; (2) not inform people that their searches were being censored; and (3) be a lot of work, with no likelihood that they could ever keep searches uncensored for very long.

By playing totally dumb, and saying "huh, we've noticed some kind of technical problem with your search, try phrasing it a different way," Google (1) takes no overt action to evade the censorship; yet (2) lets everyone know it's happening and subtly encourages them to try to evade it themselves; and (3) sets up a system that sees any censorship as damage and routes around it, which hopefully is an approach that's relatively difficult to counter without blocking the whole site, and is how the internet ought to work anyway.

In conclusion: from the little I know of what's going on here, this is a seriously clever move. Interested to see how China responds.


"""By playing totally dumb ..."""

In China, that is actually the usual way to get around the law.


It's also the authentically Chinese way to do it. http://www.npr.org/templates/story/story.php?storyId=1249130...


As everyone else has said...

This is not very new or innovative, it's pretty standard actually.


Removing the censorship was never the best way to help dissent in China. Simply indicating that the requested search had some results censored is the most potent way to fight censorship and stew dissent.

When I suggested this a few years ago and criticized Google's decision to un-censor the results, all my comments were brutally modded down.

It seems that Google has finally seen the wisdom of this approach. The next move for China is to ban Google completely, which is sadly much easier now that Baidu has had a few more years to gain traction and market share.


"It seems that Google has finally seen the wisdom of this approach. The next move for China is to ban Google completely, which is sadly much easier now that Baidu has had a few more years go gain traction and market share."

You're mistaken here. Baidu had the lion's share of the search market in China when Google decided to leave. Baidu even mocked Google's ineffectiveness with some viral videos.


My point is that if part of the goal is winning market share from Baidu the situation is worse now than before. I agree that a big part of Google's posturing about China was trying to find a graceful way to tap out w/o seeming weak.

Google's interest in China's political system is part of its marketing campaign, not actually designed to foster political dissent in China.


One of the reasons they back down on China was because they could not compete with Baidu. It's not just the ethical 'I won't censor my results' (which in my own view, is a good thing).

I'm not sure how their decision went, but I assume that they did their research first to see if their changes would really drop their marketplace share too much. I'm not sure if China will block Google... if they do I don't think it would be much different than Google censuring their own results (the censored content would not be accessible). When people try to go to Google they would notice it's censored too.


They ignore China because it was bad publicity and not worth the rounding error gain in net profits. There are plenty of obvious problems hosting a website in China but you need to add it's a poor country where it takes several times to traffic and several times the resources to make the same income.

PS: Don't believe me? in a country with 457 million Internet users... Baidu currently has a 75.8 share of China's search engine market, according to Beijing-based research firm Analysys International. Google is a distant second with a 19.2 percent share. And with that 75% share Baidu, had a net profits for the quarter ended March 31 were US$164 million. Compared to Google's $10,645 million last quarter. (http://www.pcworld.com/businesscenter/article/226507/chinas_...)


I agree with your analysis of the competitive aspect of Google's decision.

But by un-censoring the results Google removed the information about what was actually being censored, which is actually more available than the censored content itself (which, if it has any value, is likely readily available if you know what you are looking for).


I can imagine this blog post was very carefully written; Google subtly avoids mentioning that these "interruptions" are likely due to the massive censorship measures in place in China. Can anyone explain why seemingly innocuous terms like 江 (Jiāng) are being censored?


"Jiang" may refer to "Jiang Zeming" (former President some ten years ago).

There was some rumors a few weeks ago that he had died. China is currently in a transition of power from one to the next president (there have only been two such transitions so far, and there is no established legal process for it yet). Jiang Zeming is seen as a stabalizing factor during that process.

So, having rumors floating around that he died, would be potentially risky to stability (because people very easily believe stuff and easily get nuts about totally unfounded news, because you never know what to believe and what not if there is no independend press).


More detail here: Jiang Zemin death rumours spark China web crackdown - http://www.bbc.co.uk/news/world-asia-pacific-14054456

Not a few weeks ago though; appears to be from nearly 10 months back.


There's something I don't get - why is the Great Firewall of China only blocking these terms in GET requests?

If it's only GET, then a user can use POST to search and bypass all this entirely - Google just has to change the request type, which IMHO is much easier than what they just did.

If it's both GET and POST, then I don't see how typing in the infringing character which is then POST'd via AJAX is any less prone to connection termination than a GET search request.

It doesn't add up.

EDIT: downvotes? Seriously?


The great firewall is much more advanced than that.

See: http://en.wikipedia.org/wiki/Internet_censorship_in_the_Peop...

Methods

* IP Blocking

* DNS Filtering

* URL Filtering

* Packet Filtering ( TCP, FTP, POP, SMTP, ETC)

* Connection resets of up to 30 minutes

Try this: http://search.china.alibaba.com -> Search for "Falun Dafa", you'll be cutoff for about 5 minutes or so completely. That includes not just GET, POST, BUT ALL TCP/UDP connections will be rest.


It's possible google stores the finite list of bad queries in JS and has the user's browser do the comparison client-side.


Yep - the web page is a client-side app, proactively warning the user about problems in the network


Let's get that list!


I wonder if they load the whole list of "bad" terms on the client-side. Wouldn't that trigger the firewall? Unless they encrypt it, or use a bloom filter (which would save some bandwidth)...


not sure if a bloom filter with false positives would be a good idea in this case.


I didn't think of that - that certainly would be a viable solution.


You're assuming that the GET request is the one being blocked, but it looks more reasonable that it's actually the response (the html results page) that's causing the problems. Probably the "Great Firewall" detects a page that includes frequent use of some "bad" term and blocks the connection.


Maybe they are blocking POST requests too with those terms? It's not that they cannot access them on POST requests.


The same blog post in Chinese is provided as a 10 MB PDF [1]. My PDF reader simply gives up while loading it and prints insane amounts of error messages. I'm guessing they converted all the text to curves to avoid being censored?

[1] http://services.google.com/fh/files/blogs/google_chinasearch...


My over engineered solution. Catch the bad letter and replace it with a something else(i.e. another value mapped to a separate fontface). Translate the request on the server side being sure to reencode the page if it contains a 'bad letter'.

Demo: http://christopherwoodall.com/ceasar/


Well, yeah, you could also get a proper VPN that encrypts your traffic, but that's a bit beside the point.


That's pretty impressive :)


I think this is quite a brilliant move for Google. I know a lot of Chinese people considering Google inferior, because "I often don't get anything when I use Google. It must be a company of inferior technology." Actually, they have just encountered the GFW's reset, and for common people it's natural to blame the site. Now with this move, these people will gradually know that it's GFW's censorship.


I am totally naive to these cat and mouse games, but why does not Google just force SSL in China? They would have to turn off the partial searches feature, which has been shown to leak the cleartext, but I think it would totally fix any sort of censorship. Then again, maybe this type of response works better politically.


A few guesses other than the man in the middle one. One is that China might then just block Google ssl requests completely. Another is that if China did block Google ssl requests completely, it would also break any other google.com services already on SSL, such as gmail. Though of course google could host these on different domains to solve that problem.


SSL only works if you trust the certificate authorities.


Yes, but that would be likely detected, and the CA that had its private key stolen or coerced for government use would be exposed nonrepudiatably. All browser makers would immediately drop the bad certificate. In the case of targeted connections, this may avoid detection though.


You seem to be unaware that businesses (and governments) can legitimately buy keys that allow MITMing SSL connection or they could just be a CA themselves (no problem for China).

It is annoying that people downvote you instead of explaining your error in your assumption about SSL.


Yes, but then that bogus certificate is in the wild. Once once someone has a copy of a bogus certificate, then they can prove that that CA is corrupt. That CA loses its business model. What I am saying does not prevent one-off attacks, but all it takes is one person to capture a bad certificate to discredit a CA. Hence it would not work in a universal censorship scheme as Google is combating. Maybe I am still overlooking something, and I suppose China could just SSL proxy the whole country, which would defeat all of this.


You are very confused about how SSL in the context of HTTP works. Here's the best talk I know of this subject:

BlackHat USA 2011: SSL And The Future Of Authenticity: http://www.youtube.com/watch?v=Z7Wl2FW2TcA


Brilliant. Raising awareness will not let people forget and become comfortable with the status quo.


I'm in China right now and simply astounded at the apathy I see from even very well-educated folks, though. Local Chinese roommate works at Oracle and is completely fine with censored search, doesn't use VPN, etc. And this is beyond the less technical people, many of whom aren't even aware of the firewall.

I think this is still a clever and classy move for Google, though I just wonder if it motivates people beyond those already intrinsically motivated (a relatively large % of current Google.CN users).


You're astounded because well-educated people from other cultures don't care about the same things you do?


Actually, nobody will notice, because Chinese don't use google in the first place.


That blog entry was pretty non-technical. Does anyone have a clue why this might be happening? It's not like 江 is in any way uncommon.


The implication is that the Great Firewall will see a GET for google.com/search?q=[any of those characters] (or equivalent) and not let you get to google for a minute afterward.


But why would the Great Firewall block such common things?


Presumably, they're characters that are either in sensitive personal/place names, or they're part of commonly used and politically significant euphemisms or idioms.

There's also possibly a punitive element of this: in addition to whatever issues they have with some search terms, the Chinese authorities may have decided to be trigger-happy with the banning as a punishment to Google for ceasing its cooperation on search filtering.


In this specific case, the character is the last name of the former President of China and chairman of the Communist Party, Jiang Zemin(江泽民, http://en.wikipedia.org/wiki/Jiang_Zemin)


The Three Gorges Dam on the Yangtze river causes a lot of environmental damage. They dont want Chinese citizens to read about it.


As a Chinese, I think this feature is quite useful for us.Though I use VPN most of the time, I do dislike being interrupted for several minutes when I accidentally searched some censored words.


Why do people like to start a point with "As a x"? Does it mean that he/she speaks for the group as a whole or does it add extra relevancy?


extra relevance. it's a valuable datapoint to see what someone who is actually affected by the censorship thinks of google's new measure.


So everybody plz try to help us, voice of a google user from China.


And so, the wall crumbles.


Since when is McDonalds a sensitive/censored query? (And why?)


Way to beat around the bush. This is the botched censorship system, and at least Google used to make that clear. Nowadays, they show you 10 minutes of HD video of someone trying out various search queries in various browsers and being blocked.

And they have the chuzpe to add "we have checked our systems and couldn't find anything". Way to make yourself complicit.

There have already been congressional hearings into Google's business in china, back when they still tried to enable unhindered access for users there.


I did find it interesting that they got through that whole article without using the words "blocked" or "censorship". However, your read of this is completely off-base.

Reading between the lines: "Since we moved to Hong Kong and stopped filtering search, the Chinese government is just cutting off internet connections that are seen searching for sensitive terms; here's a workaround that will give you advance warning, so you can rework your search phrasing."


Actually, I think this might piss off the government more.

Imagine Google could perfectly predict what would be censored before even querying. Now you can search for related, uncensored terms and find the information you want without having to play a costly guessing game where if you guess wrong, you temporarily lose access to information. This will make evading the censorship easier by lowering the costs of exploration.


This stinks of cowardice.

If Google has reason to suspect that these interruptions are caused by active censorship by the Chinese government and are unwilling to do so out of fear of that censorship we are all in trouble.

It's not as if the population of China is unaware of censorship. Helpful reminders of what is socially acceptable and what are not have appeared in colorful cartoon form for years online.

It's not as if the Chinese government denies controlling access to information. They see it as a fundamental good to maintain the stability and health of society.

The only place and time where this matters is where the censorship is cast is a negative light. China doesn't want to be insulted on the international stage, and it doesn't want companies promoting unrest among its people. However that criticism is exactly what must happen.

Google could describe what's going on as relating to 'sensitive' terminology. They could describe the 'restrictions' as being imposed for the good of society and apologize that non-sensitive terms are being caught up in the fight for the greater good.

OR they could grow a pair.

Corporations in power have a responsibility to fight for the rights of those people without power. They have a responsibility to do so visibly to set a good example for others, and they have a responsibility to directly and honestly criticize people, companies, and governments they see as violating those rights.

This is an active debate with no guarantee of an outcome that favors American style freedom of information. If the players who can fight this fight won't, we're* fucked.

*We being those who want American style freedom's to exist globally.


They are fighting it by giving Chinese internet users access to a search engine that reveals when censorship is occurring. This is not a move that the Chinese government will be happy about. Google needs to carefully toe a line here. If they take a harder stance then the Chinese government will block them 100% from operating in China, which does far less good in fighting censorship.


The most blunt approach is not always the best.

People know how to read between the lines.


you have no appreciation for subtlety. google are pretty clearly saying "look, china is censoring these searches", but they are doing it in such a way that the chinese government can't object without looking vaguely foolish. they're both making a show of sticking to the letter of the law (after all they havent said anything other than "there's a technical issue we can't figure out, here's a workaround in the meantime") and playing to the global audience.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: