Monday, September 17, 2007

Thesis types

At the moment three out of my four PhD students are about to finish their thesis. The thesis are all very different. Mohammad Farshi's thesis is quite technical and very foucussed on geometric spanners. Damian Merrick's is foucussed on metro map layout and maybe not as technical as Mohammad's, instead he uses several different approaches. Finally, Mattias Andersson's is much more dispersed and considers many different problems and techniques. Is any of the thesis more preferable than the others?

When I was a PhD student I remember that I was a strong advocate of Mattias' type of thesis (which was also my type of thesis). A new day, a new problem. But nowadays I would probably prefer Mohammad's type of thesis. I believe that there is an added research experience in closely focussing on one type of problems. It takes a long time to really understand a hard group of problems and the satisfaction of solving a problem that you've been working on for a looong time is hard to describe if you haven't done it yourself.

Are there many other types of thesis in CS theory? Do different countries require different types?


  1. i am still in favor of "Mattias-type of theses- many problems many techniques. but it should depend mostly on the people involved, students and supervisors.

    what type of thesis would you consider better for australian standards?

    in australia phd's a bit short, 3 or 3.5 years. and that's just after a 4-year undergraduate degree. so you can get from high school to a phd in 7 years total.

    this has lots of limitations but i am not sure what type of thesis would be preferable for such rushed market-optimized resource-constrained phd "products".

  2. I agree with your comments regarding the Australian system. As a result the theses here are technically not very strong. Most theses focus on a problem and then apply different techniques. I've even heard this stated by Profs: pick a topic, apply your X standard techniques and you have your thesis.

    Another problem is that many supervisors here don't have enough knowledge about the topic that their student is working on. An exceptional student can still produce a good thesis, but I believe it's a problem for the average student.

  3. about the 3-year phd issue: there is the argument that 3 years are fine, as long as you do not publish during your phd. i am not sure if it is a good idea to go through your phd without getting some conference and paper-deadline experience.

    i think the 3 year phd is a disadvantage especially for the good students in australia. and we certainly have our fair share of good students.

  4. Publishing your work during your PhD is crucial, especially if you’re supervisor is not an expert in the area. It’s the best way to check the quality of your work.

    A major drawback for Australian students is that they can only afford to go to a couple of international conferences during their studies. Thus they do not get exposed to the international research community in the same way as students in Europe or the US. Maybe that explains Australia’s research “isolation”.

  5. This comment has been removed by the author.

  6. No offence to anyone, and I may be stating the obvious, however...

    An Australian-educated end-product is very very weak. I mean, obviously some (or many) of them are (very) smart -- but the amount of knowledge they know is soooo low.

    A person who did his/her undergrad + honours + phd in Australia would have done only 30 courses in their entire university study. Compare this with the 60+ courses done by a North American counterpart !! I mean, where is the comparison ?? The amount of knowledge carried by an Australian product is so limited/narrow.

    If I ever become the head of a department, I would be highly hesitant to hesitate hire a guy who did all his univ education in Australia -- unless the guy is super-duper like Terry Tao or Richard Karp.

    A broad knowledge about all fields also gives you a big advantage in research -- NOT just teaching. We often apply techniques and concepts from Field-1 in Field-2 in research, etc...

    The above comments also add to an earlier comment that "many supervisors in Australia don't have enough knowledge about the field their students are working on". If the supervisor did all his/her univ education in Australia, then you should expect this sad situation...

  7. Yes the situation is a bit sad. However, I want to point out that it's not the lack of good students it's the lack of a good education. The students in my computational geometry course (honours) are in general very smart students and they are surprisingly eager to learn. The problem is the education they are offered. Many students at a masters/honours level were never thought how to write a proper induction proof and some never heard about Turing machines. If you can't write an induction proof and you don't know what a Turing machine is can you have a degree in computer science?