Time2Code Level 5 4. Random Name Generator

[MattMcC-UK]

My students are having issues with this task. When they test the code, it always gives them errors. Even when we have entered the code from the completed parsons problem and made sure they match the code is always throwing up an EOFError EOF when reading a line File ":28"

Is there something we are missing?

Many thanks in advance for your help

[pdawyndt]

@MattMcC-UK I had a look at my notes for this exercise, and it indeed reports an inconsistency between the description of the assignment (in my interpretation) and the sample solution. The sample solution is also used to compile the Parsons puzzle.

In my interpretation of the assignment, only a single random name had to be generated. So the result should only consist of a single line (and no input needed to be provided). Hence the "EOF when reading a line", which is triggered when a submitted solution tries to read a line of input and there is none left (in this case no input was provided at all).

However, the sample solution (and Parsons puzzle) of the assignment assume that the multiple names must be generated (one name for each ENTER) with the input being closed by a line containing "end". I have now aligned testing submitted solutions for this assignment according to this expected behaviour, which means that now solutions have to

• output a first line containing "Press Enter to generate a new name, or input 'end' to quit."
• print a valid name (first name + space + last name with proper casing) for each ENTER in the input
• number of generated names must correspond exactly (not too many and not too few)

Could you check if this is also how you (and your students) interpreted this assignment? If so, I'll update the description of this assignment to make it more clear what the expected behaviour should be.

[MattMcC-UK]

Afternoon Peter,

You are right that the description of the problem does lead you down the solution you described. My students were using the flowcharts on the time2Code website to help with the logic and in the main flowchart main-program-52.png (1050×1002) It adds the extra lines of code that implement the end. When the program was not working and they looked at the parson problem they could see they had written the code correctly but was not passing the test. So, the instructions are not matching the problem solution from the flowcharts and parson problem.

I also think we have a similar issue with darts in level 5. We are getting another EOF error Submission - Darts - Dodona This student has changed some of the code now based on the comments generated in is_dart_valid(). I have tried submitting the solution from CnD
05-07. Darts solution.py
, and it generates an EOF error too. 

Many thanks

Matthew

[pdawyndt]

@MattMcC-UK My notes for the "level 05: darts" exercise:

• Time2Code sample solution starts from 101 (not 501 as specified in the assignment) and has a “Bust!” output that is also not specified in the assignment; also not specified is that “Shot!” must be output if the player throws 25 or 50; sample solution also doesn’t seem to take into account that player gets busted if score goes to zero but last dart is not a double (or a bull’s eye?)
• what should happen if a dart is invalid?
• not possible to make a distinction between testing with 501 or 101 if scores are hardcoded in the submitted solution (should be passed as argument(s) to functions or read from stdin)

I'll have another look into how we automatically assess this assignment, to see if some of these issues can be alleviated.

Did the updates for the automated assessment of the "level 05: random name generator" work for you? If yes, I'll finetune the description of the assignment to align it with how the assignment gets tested.

[pdawyndt]

@MattMcC-UK After closer inspection on this assignment, I think the automated assessment has been setup correctly. However, I think Craig 'n' Dave's sample solution (which is also used for the Parsons puzzle) has a minor issue. Line 50 should read as

elif score[player] - total <= 0 or score[player] - total == 1:

which could even be simplified as

elif score[player] - total <= 1:

instead of its original form (with a strict less than compared to less than or equal)

elif score[player] - total < 0 or score[player] - total == 1:

The difference is subtle, but as the assignment states a player does not win if his throws reduce the score to zero but the last dart is not a double. See for example test case #15. in turn #9 the player starts with a score of 43 and throws a double 20 (40) followed by a 3. Since the 3 is not a double, the player gets busted after two throws and should not have a third throw in this turn (as is the case in the sample solution you provided).

If you agree with this, I'll modify the Craig 'n' Dave's sample solution (and the Parsons puzzle) for this assignment.

[TomNaessens]

Hi @MattMcC-UK! We've put up a fix in addressing all issues.

The description indeed had drifted from the sample solution, the Parsons puzzle, and the automated tests. Students who followed the prose wrote a single-name program, but the test harness feeds extra Enter keystrokes to drive the Press Enter to generate a new name, or input 'end' to quit. loop — hence the EOFError the moment input() is called a second time.

What's changing:

• Random Name Generator — rewrote the "Make" / "Success Criteria" / "Typical inputs and outputs" sections to describe the interactive loop and the end sentinel. The generate_name spec was already correct and stays as-is.

• Darts — three independent drifts between the description, the two sample solutions, the Parsons puzzle, and the tests:

  • Starting score: the description said 501 (with a stale "for testing start at 101" hint), but solution.time2code.py and the Parsons puzzle initialized score = [101, 101]. A student copying from the Parsons puzzle failed every test on the very first line of output.
  • Bust! output: tested in three scenarios but never mentioned in the description.
  • Invalid dart. re-prompt: implemented in the Time2Code solution and Parsons puzzle, silently re-prompted, but was never mentioned in the description, and never tested.

  Aligned everything on 501, dropped the 101 variant, added explicit description items + I/O transcripts for Bust! and Invalid dart., fixed the example solution to print Invalid dart. on bad input, and added test cases that exercise both paths.

• Darts bust check edge case (thanks @pdawyndt for catching this on the same thread): the Craig 'n' Dave bust check was elif score[player] - total < 0 or score[player] - total == 1:, which missed the case where the running total brings a player to exactly zero on a dart that isn't a double. That case fell through both the win branch (requires a double) and the bust branch, so the turn would silently end with a phantom 0 scored. without declaring a winner. Changed to <= 0 updated the description prose, and added a test case (Treble 7 to finish 141) that actually exercises it.

Should land on Dodona shortly!