ThoughtShapes Shaping Ideas With Thought

Rjae and I are up to 2025 tests on the project we're working on together. As you can see in the image of VStudio below, they run in 25 seconds.

We recently ran through the first phase of user testing and I had roughly 45 issues assigned to me. This first phase of testing was the first time the users used the system. Out of the 45 issues, only four of them were what I think of as bugs (if you accept a bug as code that is not properly carrying out the task explicitly intended by the developer of the code). Out of these four, two were threading issues found in temporary code which, because it is temporary, I elected to have user testing uncover instead of spending an inordinate amount of time writing tests for (I decided I added more value to the project by not testing this aspect of the code because it is temporary). Each of the remaining 41 issues were a result of one of the following:

1. Functionality not implemented
2. Enhancement request
3. Change in requirements (usually resulting from the user's experience during testing)

I'd say that 95% of the issues were either a result of functionality that was not yet implemented or the result of an enhancement request. As for the core functionality of the system in terms of the client's business (I'm speaking of the domain of the business) there were no reported issues.

For those that are interested in some statistics...

It's been 48 business days since I reported having 1000 tests, which means we've done 1025 tests in that time. We've therefore been creating approximately 107 tests a week between the two of us. I know there was a period of about three weeks (and I'm being conservative) where we were at a virtual stand-still in terms of the number of tests (due to a number of reasons), so if we factor this into the analysis it would leave us with about 155 tests a week. This boils down to about 77 tests a week for each of us, which is roughly 15 tests a day. Adding in the original 1000 tests over 35 days, we have 2025 tests in 83 days.

We have 10652 lines of code and 15673 lines of test code (these figures do not include comments or blank lines). This leaves us with approximately 1.5 lines of test code for every line of code.

The system is a lot larger and more complex than 10652 lines of system code would lead one to believe. Two reasons for this are that we do not have any generated code (which often pads code-count statistics) and we have been very attentive to the DRY principle (Don't Repeat Yourself). Why no generated code? Hmmm...I think that is a post for another day.