Via slate.com from unknown source

So this is a study on 300 Asian families compare to Amy Chua's sample size of 1. Sorry, despite Chua's portrait of relentless strive for excellence, the study finds this actually result in lower academic achievement. Study also finds "Tiger" parenting is not typical among Chinese as in her stereotypical portrait either.

Perhaps less surprising is the kids show more depressive symptom. I'm looking forward for Chua's children to write their own book to tell their side of story!

Next I came upon the old observatory now restored into an upscale shopping complex called Heritage 1881. Tsim Sha Tsui is becoming more glamorous than ever.

Start Ferry is my next activity. I was just going to ride the Star Ferry to the Central and come right back. But I found the pier in Central is also changing and is becoming a destination itself. There shops and restaurant and a farmers market. I found a newly opened Maritime Museum there and spent two hours to check it out. It is good to be a tourist for a day.

I found these tickets in the bottom of the drawer, yellowed paper tickets dated December, 1999. It was from the one trip to the United States my father made in 1999 for our wedding ceremony. After we went back to work, he and my mom would venture out on their own, using the bus map to find their way around the city. He did not stay long through. After a little more than a month, he returned home. These tickets were keep as his memento.

The movie is 不老騎士 "Go Grandriders". A group of elderlies, average over 80s, are out to accomplish a dream. They want to make a trip around the island of Taiwan on motorcycle. This is a physically demanding trip that last 2 weeks! Despite of their aged and frail body, the granddads are determined to do this. The documentary follows them along. There are challenges, there are assorted health problem to deal with, a lot of tears have shedded (because of their determination to tackle the problems). It brings a lot of happiness to the granddads when they finally accomplished their dream at the end. This is quite an inspiring tour.

三月五日清涼的深夜，我和家姐步出佛教醫院走到外面無人的街頭。五年前媽媽在這醫院過世，這夜榮業也離開我們了。榮業你再不用受痛苦的煎熬了，安息吧，我們會懷念你的。

To honor my father, I ordered his favorite BBQ Pork Rice in the same food court that we have gone to.

Having to develop and maintain your own rendering engine and be compatible in such environment is difficult. I hope this move will allow Opera to live on.

Pixar make this an exciting adventure story from the point of view of the toys. They are masterful in making a sentimental story out of toys. At then end Andy give the toys to little girl next door. He reminisce the old time with the toys once more before passing them on.

In the movie, the solution is to pass on the old toys to people who want them. In real life, the problem is of a much bigger proportion. In the modern materialistic society, we have acquired so much stuff that it is causing much problem due to the clutter and the constant battle to keep the home organized. We face a similar issue on decision of the disposal of the toy after it has passed its days or simply because we run out of room. Giving them away doesn't solve the problem because there a lot more than people want to receive. Nor is throw them away easy because of our horder nature.

If you think Pixar made a movie to show us what we should do to toys as a responsible owner, look at what I've got and think again. In the first few years of my sons' life we have collected a whole parade of Toy Story characters, many of them are gifts from others. Pixar is a for profit company. One of the way for them to make money is to sell as many Toy Story merchandises as possible. So we end up with not one but three Buzz Light Year of various sizes. In real life, as oppose to in the movie, Pixar is very much a source of problem.

昨天我去洗澡，留他一個人在廳，回來看見他，他對我說：「搞」。

這樣一個「搞」字我覺得怪好笑。搞什麼？我一時沒有看見，因為要忙替他洗澡所以也沒有再留意。後來才發現原來他拿了我的銀包來玩，把裏面的東西拿了出來。總算他自己坦白招供。

今早他說得更詳盡，我步出家門一小會。回來時他跑來對我說：「搞搞震，行行曳」，原來是把哥哥的玩具稍微弄壞了一點。難得他自動招供認罪自知是曳，小孩果然天真無邪。

Hindsight blinds because the path of what happened in the past is so brightly lit. That trail is so clear to you that everything else that could have happened is cast even more deeply into shadow. If you then want to reconstruct what could have happened and understand the choices made that produced that pathway versus another, you actually have to work hard to avoid the dangers of hindsight to then re-imagine the world before problems had been solved. — Philip Zelikow

• 抹乾淨眼
• 兩個波
• 紙包蛋糕
• 哥哥食嘢包包

開始懷念他牙牙學語的時候了。

Coursera has finally issued me the statement of accomplishment for the Probabilistic Graphical Models Course I took last December. Since 2012, I have completed 7 online courses with Coursera. I have the faintest idea about the subject when I sign up for it. Little do I know what a killer course it will be.

So what is Probabilistic Graphical Models about? It is a system to model the world with lot of uncertainties using probability and network. The application include medical diagnosis, traffic prediction, speech recognition and so on. It requires lot of proficiency in math and probability. This is really pushing the limit of my knowledge. I have spent hours pounding on the programming exercise. It actually foster a lively online student community to tackle these really tough problems together. This completion statement is definitely the most hard earned one. Just don't ask me how much I still remember after the course.

Along with Probabilistic Graphical Models, I did the seemingly suicidal feat of taking 2 other courses at the same time. It does take a lot of effort to keep track of three courses. But in terms of workload, PGM is so overwhelming that those other courses does not seem to add much overall. In fact they sometimes make a nice diversion to the PGM course.

I stopped to chat with a fisherman. He was pulling up his net and picking up a few fishes into his bucket, which is already half full. He told me it is the herring season. For a few days a year large number of herring are running into the bay. The fishermen have a good catch and the birds have a feast. It was quite a sight.

The main reason for not having completed the book is not because it is not interesting or not useful. It is very much an excellent book to introduce readers to the topic of collective intelligence. The problem is I strive to work through all the coding examples and not just to read the text. Many times I have feel the examples given, designed for average programmer without additional experience, often has low abstraction level. I tried to improve them, including building some tools on myself. This unfortunately doomed me because I never have the time to build the tools or even just to work through the exercises in sufficient detail. This time around I feel that I have gained a lot of experience and have access to many tools like Pandas. I hope this will be an enjoyable exercise.

But if you look closer at the result, a recount should not be more costly than a plurality voting election. The fact is this is a race between F.X. Crowley and Norman Yee. No other candidate come close. From the detail D7 Results Report we can see that the second runner up Michael Garcia was eliminate in round 5. His has 7473 vote, trailing the first runner up F.X. Crowley's 9709 vote by 30%. He could not have possibly make it to the last round. This mean the race is between F.X. Crowley and Norman Yee only. No other candidates will have nearly enough vote to make it to the final round.

In this case we can adopt a simple recount procedure. Instead of running a full ranking, we can accelerate this and eliminate all other candidates right away. If a vote has both Yee and Crowley on it, it should goes to the higher ranked candidate. If it has only Yee, irrespective what rank it is, it is a vote for Yee. And similarly a vote has only Crowley goes to him irrespective of what rank it is. If a vote that has neither Yee or Crowley it goes to the exhausted ballot bin. This allow us to do a quick recount without painstakingly ranking every choice. And the result will the same as doing the full ranking.

It may be too late for San Francisco to adopt a new recount procedure after the fact. But if both candidate agree this is a fair way to do it, I think they can push this as a preliminary recount procedure. The incremental cost of a simple recount should not be much higher compare to plurality voting. And this should help settle the doubt of the election result.