On Jan 2, 2015, at 10:08 PM, Darren Cruse wrote: > a. Testing > > Regarding my poor attitude towards TDD... :) I have been trying to take > everyone's comments to heart and have been attempting to get that black > background stuff off of drracket for my pong program!!
HtDP is NOT TDD. (I learned 'testing' in 1984 in a Franz Lisp and Scheme system. TDD is just the popularizers' way of getting a small element of good programming out to the masses.) The tests/examples serve as part of the design step. How can you solve a problem w/o making examples? So HtDP is trying to create a habit. > I have made some progress but I did want to ask a few pragmatic questions: > > a1. Do people *truly* go for 100% coverage i.e. *truly* get rid of all the > black background stuff? No. Students are definitely allowed not to test the main function of a big-bang program. > a2. One kind of funny question - I attempted to add a test of my main > function - which calls big-bang - esp. since drracket's showing the black > background on that. Something like this: > > (check-expect (main initial-world) expected-world) > > (define (main world) > (big-bang world > [name "Pong World"] > [on-tick handle-tick] > [to-draw draw-pong-world] > [on-pad handle-key-down] > [on-release handle-key-up] > [on-mouse handle-mouse] > [stop-when quitting? draw-goodbye])) If you wish to test such functions, and I need to do that, add another argument to main that specifies for how many ticks you want to run the program. Yes, I should create a big-bang clause [testing? Boolean] so that people can test on a head-less machine, like we do for drdr. > a3. I also have quite a few of these "setter"/"updater" type of functions > for copying a structure but setting one of it's values. Here's one for an > example: [It's "its" not "it's" here.] Again, I'd expect students to test a function that uses pong-set-world-set-ball (what a name, how about set-pong-world-ball). In general, if you are designing f and you put g on the wish list for f, I would expect that you at most derive an example from f's suite for g's -- simply for the design of g. If it is obvious, I would say, just define the function. If f fails and g is even remotely a candidate, derive tests for g from f's test suite to rule out mistakes in g. This is for freshman students who have matured to a certain points. I think your pong program demonstrates that you have gotten to that point. Why don't you tackle a universe-based pong game? -- Matthias
____________________ Racket Users list: http://lists.racket-lang.org/users
