FreeGameDev

I've gotten a lot of comments about the lack of working unit tests for my project, and also the difficulty of installing the server. I'm wondering, if you, as a developer would ever consider working on a project that does not already have working unit tests in place, if so, under what circumstances would you consider it.

I have to admit the bar is pretty low for me to want to work on projects, simply this,
1) compiles and runs on my system (rules out (98% of projects)
2) isn't so far advanced that any other contributer would make a better contributor than me (probably the other 2-1%)

I think the need for unit tests depends on what kind of software it is. For example, they make perfect sense for compilers, but they are not very useful for word processors or games (imho).

When you say "lack of working unit tests", are there any unit tests that are not working? That can be a bigger problem than not having any unit testing at all.

Yes I would consider and I would work on a project that has no unit testing at all, though I really would prefer if unit testing was already practiced by the project's developers. I've worked on games without any unit testing, or testing besides letting users play the game, and games with unit testing enforced, the later are really better to work with, even if you don't actually understand much about unit testing just the simple ease that having an automated test grants you by checking if you're changes have any unforeseen side effects in the rest of the code is useful.

Though I do understand that for a project with a very small team, having to write the additional code for unit testing may pose as a waste of time, and it may even be, though it is harder to add unit testing to existing code and takes much more time and effort than to do it while you're writing the code, the "unit test before code" idea is important.

Yes, unfortunately, the project has unit tests, and some of the unit tests no longer work. Basically, the project is a fork of another project, a project that well, went in a direction I felt did not have the appeal (for me) of the direction I really wanted the project to go in. The project was changed from a top down view with square tiles to a 45 degree isometric. Some of the archetypes were changed. the client/server connection protocol had to be changed.

The old client java unit tests only worked under one ide, and we found ourselves using a different ide.

The project team is tiny, only one programmer, really, and only two active contributors (other one is a tester/documenter artist)

Of course if having working unit test is necessary to get more contributors, it becomes a chicken and egg problem.

I would work on a game project that does not have unit tests.
Then, I might want to see if I can get unit tests working, as well as find and remove bugs in the code through
whatever other means.

Jastiv {l Wrote}:Yes, unfortunately, the project has unit tests, and some of the unit tests no longer work.

Then I would suggest removing the unit testing code unless you can find someone to maintain them, and given what you've said the existing tests would take as much time to update and check as creating them from scratch. If the code does nothing and is not working, worse if the it doesn't help you, than just remove it and prevent any more confusion to possible contributors that may think the tests are supposed to be respected. If you're developing the project and are not using unit tests then there is really no point in keeping them.

Knitter {l Wrote}:If you're developing the project and are not using unit tests then there is really no point in keeping them.

Seconded.

I think the lack of unit tests will not deter potential contributors -- if you are the programming guy then the most important kind of contributor you can get is artists (and musicians etc) and they won't care at all about unit tests.