Google keeps on losing market share

Google keeps on losing market share to the domestic search engine Baidu, reports China Web2.0 review based on a new report by CNNIC. Baidu now get 74.5% of the Chinese internet users as primary users of their search engine, compared to 62% last year.
That can be explained by the fast growth of the internet by 30 percent over last year, when new Chinese users will most likely use a local heavy weight in stead of a foreign less known search engine. More troublesome is that Baidu is also winning in the high-end users. Google get 22.11 percent of the high-end users in the first-tier cities, but only 5 percent in the third tier cities.

Internet nanny drives me against the wall

For a while I thought I could live with the internet censorship, or our nanny as some call the grumpy old woman. While being irrational and wrong in itself, I would be able to find patterns I was sometimes even able to explain to outsiders.
For example, for a long time - apart from real emergencies - new IP-blocks would only get in place at the first working day of the month. This was all based on the basis principle of any bureaucracy: we take ourselves very serious, but do not like to work too hard.
By reducing the actual number of IP-blocks the Great Fire Wall (GFW) even became slight more efficient, since many newbies did not see a good reason to educated themselves on circumventing that wall.
But now the old woman has gone crazy and all my old certainties seem out of the window. First, she blocked Feedburner. What sense does it make to block the world's largest producer of RSS-feeds? Maybe nanny wanted to punish Google, who bought Feedburner recently, for not joining the non-sensical declaration on self-discipline on the internet, that was signed for the first time by Yahoo and Microsoft.
Now, today the English section of Wikipedia was closed again, and act that made really no sense to me. The problem of course when you cannot make sense out of it (unlike in my first example), the punishment does not make any sense. It only forces more angry users to get their proxies in place again.

Google sees censorship as trade barrier

Google offers an interesting twist in the censorship dilemma, as it asked US trade officials to treat internet censorship as an international trade barrier, writes AP. Not only it the proposal interesting, it also seems that Google is now actively engaging into governmental relationships, a field it would try to avoid in the past.
This is the argument:

Google sees the dramatic increase in government Net censorship, particularly in Asia and the Middle East, as a potential threat to its advertising-driven business model, and wants government officials to consider the issue in economic, rather than just political, terms.

Google and Sina team up

The battle of the giants is getting into place now Google and Sina have announced a strategic partnership, say reports all over the internet. The question is now, who is going to be the biggest, Sina, Baidu or QQ, and does it matter. Interesting times.

Tencent goes into search

The succesful Shenzhen-based IT-company Tencent is now moving into the search business, reports China Tech News. The company is said to have bought in April the domain name wenwen.com ("Ask a question") for 80,000 rmb.
Tencent started off as an IM-company with its popular QQ and has used its over 200 million user accounts to successfully take on weblog hosting, video hosting and now moves into search, where Baidu and Google are still the number one or two.

Update: China Web2.0 review has more details.

Search engine conference becomes Yahoo-event

The second China conference on search engine strategies, at the end of this month in Xiamen, promises to be even less of an event than the 2006 SES in Nanjing last year. Last year China's leading search engine Baidu was not present at the stage.
Yahoo-China has now become the major sponsor, while Microsoft and Google have disappeared as a sponsor and on the current agenda I could only find a Google engineer who would be speaking. Also local players are no longer sponsoring. Of course, Baidu is certainly not present on the stage.
Even Alibaba's Jack Ma is no longer giving a key note speech like he did last year, as far as I could see on the website. Probably Mr. Ma does not feel to happy with the only unsuccessful venture he is heading. While the talk of a possible IPO of Alibaba has become stronger, Yahoo China is excluded from this exercise and probably for good reasons. The conference might still be of interest as an expert-meeting, but no longer as a media event.

Troublesome: Losing email on my Gmail-account

Over the part two days it happened at least three times to me: recent emails at my Gmail account could not be found. Last night I had to call back a potential client in Florida, I routinely looked up her initial email and my reply. Both were gone. Fortunately, her contact details were saved in the contact list, at least it indicated it was not me losing starting to make things up, but that there really has been an email.

This morning I had to participate in a radio program on press freedom and I tried to find back the initial email and my subject proposals. Both were gone. That is potentially troublesome when an email service says I never have to throw email aways, and they start losing them. I'm not even a CEO of a listed company that has to lose emails.

Google sorry for stealing

Google in China has apologized to Sohu.com for 'lending' the database of Sogou, its input system for pinyin. Google offered their excuses just ahead of a press conference by Sohu where it wanted to vilify its competitor for stealing. The ongoing discussion will get an interesting twist here.

Has Google Pinyin been stolen from a competitor?

A debate is emerging on the internet whether Google has for its newly released pinyin input system stolen the vocabulary of its competitor Sogou. When the US search engine would have committed such a gross infringement of the intellectual property of a competitor, even though it would have been done by local staff, that would put the company in a difficult position.
There are two different takes up to now. Yee and others point at awkward similar mistakes Google seems to have copied from the vocabulary of Sogou. The fact that Google has failed to respond to the accusations is already seen as an admission of guild.

China Web2.0 Review is an authority here and writes:

My take is that further discussion on this similarity case is not necessary. If Google really infringes Sogou’s intellectual property rights, they would sue Google. Is it possible that they all licensed vocabulary library from the same source? You will never know it.

I tend to disagree with both at this stage. The pinyin system seem to be much more than a vocabulary and you would have to compare more than only the vocabulary to substantiate the accusation - although Google does seem to have a problem here.
The fact that Google did not react at all is of course not good, but if they would have denied guild after the first rumors would have emerged, you would know they would not take the accusation serious at all. Google needs to have a thorough internal investigation, and they could have told us that.
I also do not agree with the argument that because Sogou has not started a court case yet, there is probably no case. Also Sogou needs to investigate the case and the systems is only out. Preparing a court case takes a long time and does not yet indicate that Sogou has already taken a position here.

Google, Yahoo reject shareholders motion against censorship

Both Yahoo and Google are advising their shareholders to vote against a resolution of the New York State Pension Fund that call for an end to their corporate censorhips, reports RConversation.
Both companies do not give a reason for their negative advise and refuse to talk to the media about the issue. The annual shareholders meetings are to be held on May 10 (Google) and June (Yahoo).

Google is learning Chinese

Google has released its new way of using pinyin to write Chinese on computers and one sentence in the (short) review by Shanghaiist made me thinking.

If an user simultaneously logs into his Google account while typing, GP will analyze and learn his word entry habit, and overtime optimize character output to save user time. How cool is that?

That is not only very cool but also very smart. While my verdict on their newly released translation tool was not very mild, the build-in ability of the system to learn was promising for the future. It is similar for its book-scanning projects. It looks like Google has started a rather intensive Chinese class, helped by millions of users. (The same of course goes for all the other languages, but Chinese is of course a special challenge.) It would be interesting when somebody would lift the lit a bit of this interesting operation.

Google translation tool works only in a few cases

After last week's initial enthusiasm about the new Google translation tool, I decided to ask different people to check how the translation works out in their languages. I was able to investigate 14 different translations and the results do not look that good. For languages I do not speak I got help from my audience, most of the translations into English I checked myself.
Perhaps Google offers a huge improvement compared to for example Altavista's Babelfish, but in most cases the translations are not really usable.
What goes reasonably well, is translating Chinese into English, but the other way around is a disaster, for reasons mentioned here. Chinese has too many homonyms and Google gets only 50 percent right - and you have to guess what fifty percent.
The only good score - in both directions - is for Russian, other languages do poorly.
Philippe: "you have to guess from the context what it actually is supposed to say". German into English sucks, I concluded after trying to translated a chapter of my book on China into English.
Joerg Kilian thinks the service also sucks in German and retranslated a German translation into English of my entry on Chinese airplanes:

Not in the foreseeable future, I think. The condition gremium of China, the highest administrative organ, announced that the country would be ready for a larger commercial area (surface) for at least 150 passengers, rainsing the reported media.
After five decades development things were ready the official newsagency Xinhua reported. If I would be responsable I would let them try something more harmless like movable telephones.

I think it is the general problem. You could make sense out of it, when you know the original. But that defies the need to use a translation tool.
Mariab on Italian:

the translation of your blog in ITALIAN is quite inaccurate, actually in some parts is not really understandable. I think it works only for simple phrases with a quite linear structure.

Juan Pablo Cardenal:

translation into Spanish not good. Very confusing.

My advise: do not use it when translation is really important, the chance to create misunderstandings is just too huge. It reminds me of several situations where translators were not too sure about their own skills and just prayed nobody of the delegations they were helping would speak both languages. They are very good in holding up appearances. It is partly psychological: you want others to understand you, so you all too easy assume they do so.

Google China tries to talk to bloggers

China Web2.0 Review points at an exchange that emerged as Google China tried to deal with webloggers that were mostly negatively reviewing a new China-service called Daohang. What happened was that Google's PR-agent in China, Ogilvy, started to reply on those comments by sending those webloggers the press release on that service.
Let's first look at it from a positive side: Ogilvy did react on the comments and that is already much more than what happens mostly. But apart from that almost everything went wrong.
As China Web2.0 Reviews also says, weblogs are conversational media, you have to build up a conversation. Sending a press release is not the way to build up a conversation. So, that means, leaving comments and actually reacting on what the weblogger is saying.
Also, you have to find a real person to react, a company cannot do that anonymously. And, I would add myself, Google China cannot outsource this kind of work to a PR-company. PR-companies might be able to help and give advise, but the person talking to the webloggers should be involved in the work directly, coming from Google China.

Translations: German to English sucks

Thanks for all the response that is already coming in on my request for an evaluation of the Google translation tool I posted here this morning. As expected, people think very differently about the performance up to now. After the weekend, I will give a decent overview, since some of the feedback is also coming in on email.
Most of the feedback is on the European languages, and I do hope to get some on Japanese and Korean too. Do forward my request to others, if you think they might be able to help.
Meanwhile, I could not resist the temptation and have tried to translate a chapter of my book in German into English. Here is the English translation and here the German original, so you might be able to compare yourself.
As you might see for yourself: the English sucks. There is really too much guesswork needed.

Google sends translators home, or not?

In December I made fun about the online translation service Babelfish was offering at the time, to the relief of a few friends who make a living in translating. I thought it would take another year before I would give such a service a serious chance.
But today I got an enthusiastic email from Maria Trombly, who advised to have a look at the new online translation service of Google itself. As a picture you see my weblog in Chinese, and when you click here you get the latest update translated.
Compared to the Babelfish efforts, the Google translations actually make some sense, although I must agree with Maria:

it read like a Xinhua news report

And that is not necessarily a compliment. The problem is of course, that when the original does not make any sense in your eyes, a translations is also a problem.
Of the English to Chinese translations, my Chinese colleagues say about 60 percent is correct. The software picks the most commonly used characters, and that is not always the right one. It might be an interesting experiment. The system allows you to suggest corrections and when people start doing so, it might become better over time.
I found it a useful barrier, not being able to read too much Chinese, because when those gates open, I will be lost. Just see what happens.