View Clarifications


[ Contest Page | Scoreboard | Submissions | Clarifications | Help | Log In ]


ProblemClarification
General I/O?

Please use standard in and standard out!

Input --
Java: Scanner in = new Scanner(new BufferedReader(new InputStreamReader(System.in)));
C: scanf("%d %d", &integer1, &integer2);
C++: int x; cin >> x;

Output --
Java: System.out.println("asdf");
C: printf("%d %d", integer1, integer2);
C++: int x = 5; cout << x << endl;
General Java class names?

Please use the following class names (make sure classes are public):

Problem A - public class A { ... }
Problem B - public class B { ... }
. . .
I - Image Recognition Checker?

Not ready yet. So problem I won't be judged for now.

No this is not a message from the last week :)
General What happened to the scoreboard and our submissions?

Added scoreboard, tried to rejudge everything ,did not quite work (some concurrency issues) so I had to rejudge problem by problem.
General What's wrong with the scoreboard and why do all my submissions revert to "Not yet judged" every now and then?

Added scoreboard, tried to rejudge everything ,did not quite work (some concurrency issues) so I had to rejudge problem by problem.

Caught up.
F - Four Points Checker?

Was broken for a while - should be OK now (stupid i's and j's...).

Anyway - have you tested against the second sample case?
F - Four Points Why is our F failing?

Read the problem statement (i.e. it does not pass the second sample)
A - Auxiliary Question of the Universe I'm sorry that I keep accidentally submitting for A.


F - Four Points ... and that I keep submitting buggy code for F when I should obviously check all the sample inputs.

.. and I am sorry for the buggy F checker (found another one...).
General Are you sure my C times out?

Yes.
Judge's solution (in Java) runs in less than a second on the largest input. Our time limit was set to 5 secs. Your solution runs in 12 secs. If I had this situation in a contest, I would probably pass it (just for the lack of testing) but here it is - you have to figure out something more efficient.
General Does the checker for F check if the outputted points are in a different rotated order? E.g. {0, 0}, {1, 0}, {1, 1}, {0, 1} can also be outputted as {1,0}, {1,1}, {0,1}, {0,0}

Order does not matter, as long as it is either CW/CCW
F - Four Points Also... I've just submitted one of the the judge solutions for F. It got WA. :(

Seems that checker is wrong, sorry about that - will try to fix and rejudge.

[EDIT] Cannot SSH into our server - seems like my account has been disabled, it may take a while to fix this.
F - Four Points Thanks!

Fixed the checker - only judge's solution passes. Your 26th submission was the closest one to passing (the slow one, still under time limit).

Error in checker? abs(cross(v1,v2)) < 1e-6, although points were within 1e-6, when you multiply, error is > 1e-6. Set that check to 1e-3 and rejudged all F's. The result is what you see. Not pretty :)

The first test your #26 fails is:
-707 -246
-794 66
-990 396
-597 491

There is no solution, you output some points.
F - Four Points Thats strange, because I submitted two different judge solutions, one which outputted points in a different rotated order.


F - Four Points I will try again, might be an error on my end.


F - Four Points Just resubmitted the judge solutions, seems none of the two I submitted pass. I also submitted my solution along with the two (doesn't pass as usual :( )

I also submitted one of judge's solutions, it passes (yp). So does gk.
F - Four Points sk and the two other yp's (yp_alt) fail, but yp passes. It's probably more accurate than my code. I'll keep trying.

It is weird, but I would not spend too much time on finding out why it failed (who knows what happened during the contest). Just make sure you understand what the solution is. They have slides in Russian but picures should be enough.

My checker compares lengths of 4 sides (in order) and 2 diagonals (make sure it's square). And then that each input point is contained in exactly one side.

Meh, I hate precision problems...
F - Four Points If the points have error 1e-6, then the diagonal vectors would have error 2e-6 (from subtracting the two diagonal points). From that, the cross product would have 2e-6 * (magnitude of the vector) in error. Perhaps instead of checking absolute error, relative error would be better. Anyway, I'll just submit two-three more times and give up. Thanks for all your help though!




Ultra Cool Programming Contest Control Centre v1.8
Copyright (c) 2005-2010 by Sonny Chan