How to start a PhD

At the time of writing this, I am a bit more than 5 years after finishing my PhD. Looking back, it is one of my best experiences.

Is research right for you?

Computer science research is not particularly easy:

It requires a lot of hard work and persistency. From time to time, you need to work beyond just 9 to 5 and on the weekends. Working on a problem continuously for 3 to 4 years can't be achieved without persistency.
It needs a lot of creativity. You need to have better ideas than all the people in the world have tried, and you need to keep doing this for about 30 years through out your career.
Good people skills. You need to interact with your supervisors and senior researchers, your peers (other young researchers), and your students (when you start teaching). The outcomes of these will have a large impact on whether research is enjoyable to you.

If you decided research is not really for you, it is not a big deal. There are so many other professions your can choose from, and this is just one step in your pursuit of the ideal career for you.

Research is a great life style!

Research is probably the ideal career if you enjoy it:

Happiness. It is something you enjoy, remember? Happiness should be the number one criteria for an ideal career.
Change the world. You will have the opportunity to work on the problems that have a big impact on our world: cancer, energy crisis, poverty, water shortage, etc. You will be part of a large global team tackling these problems.
You are the boss. Researchers have great freedom of what research to do and how to do it. Also, you'll have access to increasingly amount of fundings for the research you are interested in, and you don't need to worry about the chores such as accounting, finance, legal, human resource, administration, etc (only a little :-).
Work and life balance. Academia is one of the most flexible professions I know. You choose to start working at 10am, 12pm, or just work from home (when not teaching :-). Most universities and research institutes provide very generous leave and family support.
Travel and friends. The open nature of research means you will have lots of opportunities to travel around world, visiting many exotic places (in the name of work :-). Also, you'll meet many like-minded, smart, and gentle people who share the same reserach passion as you.

How to find a good supervisor?

The single most important thing to start a research career is to find a good supervisor, but how?

Find the area that excites you. The key is to be imaginative: if you are into computer games, there are computer graphics, human-computer interaction, computer game research, artificial intelligence, robotics, and many other related fields. There will be one that excites you.
Find the good researchers in the area. This is not easy and it will you take some time. Start with the top conferences and journals in the field. You'll begin to have some idea after reading some papers. It is also worthwhile to find out what are the top research groups in the field. The most important is that their research excites you.
Find a good supervisor (that will accept you).
- It is like finding a job: you have to keep trying and knows how to make yourself stands out. Do be discouraged if your first a few tries don't succeed.
- There are many practical issues here, such as the financial support and whether the professor knows you or not. It is not fair play, really.
- Be aware of professors with too many PhD students. A good supervisor usually does not have more than 4 or 5 students, otherwise he or she just doesn't have enough time for each one.
- Also, famous professors may not be the best supervisors: they usually have lots of other duties besides teaching, such as management, funding application, various reviews, etc. This will limit the amount of support he or she can give you.
- There are many other issues such as culture and personality difference. It is not possible to know everything before hand, so trust your feelings. Life is an adventure, isn't it?