Re: I want to contribute and chose right project for graduate thesis

[Emanuel Berg]

Mario Krajačić wrote:

> All your mails were very helpful. My mentor
> is advised me to search unit testing project.
> That will be enough for thesis.
>
> Could you please advice me some unit test
> projects in C.

Testing is by now a whole research body so this
project has the advantage there should be
a couple of big books to read.

Unit testing is basically small wrappers around
library functions that do basic computation
(the lib funs do that, that is).
Whenever something is changed in the library,
unit testing is done to see if it still adds
up. So if you have a library to compute the
square of x, whenever anything is changed, unit
testing checks if x² is still 4 for x = 2.

For more advanced examples it is not
a bulletproof method, but it still should catch
most errors.

The "problem" with unit testing, and actually
all testing, is no matter what there is always
another superior method to test, and that is to
*use* the software. And if many people do it,
the better. And I'm not talking "beta testing"
here. Just use the real deal!

-- 
underground experts united
http://user.it.uu.se/~embe8573