18 September 2010

Typo in 17 September quiz leads to rescoring (1385)

Well, no excuse or explanation for this one beyond sheer sloppiness. One of the choices, scored as (and very much intended to be) correct, was this:
BEGIN
  SELECT my_sequence.NEXTVAL into l_value
    FROM SYS.dual;
END;
Sadly, I lopped off the declaration section in my copy/paste operation, and so this block of code would not execute successfully. I will, therefore, take the following actions: 1. Give everyone credit for a correct selection on this choice. 2. Fix the choice so that it includes the declration section. 3. Re-score and re-rank. Vandu wins an O''Reilly Media ebook for noticing and reporting this mistake. My apologies, Steven

7 comments:

  1. Steve,

    why rescoring? It was clearly incorrect choice because of missing declaration of the l_value variable. For me it was nice example of test of attention. For example, I ticked a choice with l_value declared as DATE type because I didn't pay enougn attention - my mistake, incorrect choice. I could say - oh, Steven, typo, by mistake you used DATE instead of NUMBER.

    Even if I know from your explanation that the missing declaration was typo because of copy/paste operation - for me as a contestant this choice is the same "incident" as "l_value DATE".

    Regards
    Ludo

    ReplyDelete
  2. I thought it was deliberately left off :)

    ReplyDelete
  3. I also thought it was wrong choice "by choice" and not "by copy and paste mistake".
    It was so evident that I didn't think there was any ambiguity.
    I really don't think a rerank is necessary.

    Thanks for the sustained efforts in giving us the challenge.

    ReplyDelete
  4. Zep - you make very strong points. When I have typos or mistakes in the quiz, it can lead to all sorts of "distortions" in the scoring. You might puzzle over that typo for a long time, which could reduce your score. Yet I do not want to punish people who do not notice MY mistake. The best solution here is to avoid the need to do this re-scoring. This past week was a bad one. But we had been getting much better and I will redouble my efforts to provide "clean" quizzes. You also suggest highlighting changes that occur to questions. I plan to do this, but I am not sure when it will be available.

    Jeff and Sterol: I could have taken the position that it was OBVIOUSLY wrong and even intended that way, so no need to rescore. The problem with this approach is that I feel strongly that a quiz should test your knowledge about the topic of the quiz, not about finding things like missing declarations. If the lack of the declaration was germaine to the quiz topic, I would not rescore.

    Thanks for all your feedback and your patience with my sloppiness.

    SF

    ReplyDelete
  5. Steven,

    You state that if the problem with a question is "germaine to the quiz topic" then you will not rescore due to a quiz "mistake". If this is the case, perhaps you could consider displaying the quiz topic when users are taking the quiz.

    ReplyDelete
  6. I have thought of that, Ryan. The only problem is that in a number of quizzes, giving you the topic can be a substantial clue in determining the right answer.

    In any case, a mistake is a mistake. It should not be there and should be irrelevant to the topic for the quiz.

    ReplyDelete
  7. I agree with the first couple of comments. I took it all at face value, checked the facts and thought - this would not compile, so I didn't select it.

    Perhaps that's a key assumption to be made - unless stated otherwise, the code should be compilable - look out for runtime issues.

    My morning thoughts assure me there have been quizzes however that do look for compilation issues, but I guess if that's clear in the line of questioning.

    Like Steven says, testing PL/SQL knowledge, not typos, but that's part & parcel of being a programmer.

    ReplyDelete