refactoring topic for undergraduate thesis?

[Simon Xu]

I am looking for thesis topics for Computer Science students? Can you gibe your opnions? Thanks.

[HS Thomas]

Are you a teacher or student ?
regards

[Ilja Preuss]

Don't know wether they make a good thesis, but here are some interesting topics:
- refactoring of C++ code
- refactoring of non-OO (functional, declarative etc.) code
- refactoring of legacy code
- when doing evolutionary design (a la XP), how likely will you "refactor yourself into a corner"

[Simon Xu]

Thanks. I am a teacher. As I am not very familiar with this area, but I think there are some interested topics.

[Mark Herschberg]

The concept of an undergraduate thesis seems to vary greatly these days. Some schools pretty much let you just do a coding project. Others want actual academic advancement.

There are plenty of refactoring tools. A simple thesis could be to build one with new functionality, or add existing functionality to an open source one.
A more interesting topic might be to develop heuristics for when and how to refactor. This could be done by theory or by empiracal study.
A thesis on the usage and cost-benefits of refactoring could prove interesting, if the student is more interested in project management issues.
Not much work has been done on large scale refactorings.
It might be interesting to consider refactoring in conjunction with porting.
If you want to get really theoretical, you could describe refactoring in terms of some abstract modeling schema. Here I'm thinking of some type of lamba calculus representation, and then describing the refactorings as transform functions.

If I think of more topics, I'll let you know.
--Mark
[ August 17, 2003: Message edited by: Mark Herschberg ]

[Simon Xu]

Mark,
Thanks for your points. Your posts were very helpful.

[Mark Herschberg]

I'm happy to help. I'd love to hear what is choosen as the topic. If you need any more help, don't hesitate to contact me.
--Mark

[Ilja Preuss]

Another interesting topic which just came up in this thread: refactoring of data models