OJ Board

There are inputs with many parents for some nodes.
And input is
you had a topic on 115 with date Sat Nov 01, 2003 2:13 am,
it is very helpfull

ball uses unit of lifetime for each HIT of bumper,
even if it does not move, it can lose all life just by hitting walls.
answer to your input is
0
0

it is couple of hundreds and, as i said, this does not causes overflow problems with this algorithm. check this: no incorrect rounding in second loop and 1 output number computation - can be done with integer division m must be >= 2 after you check b != 1 check both input numbers are satisfied dont ...

i suppose so:
loop to find m and height, 2 solutions are ok
then just second loop to find height of stack

ok, ok, i sent my solution couple of times and it seems that
max height OF STACK is >= 10 and <= 100 millions

in my solution i also had to change int to unsigned because of overflows,
then got accepted.
check boundary conditions, don't count to much in first loop,
you probably need 2 x if( ... ) break;
haven't noticed any other problems or tricks

i lost couple of hours, but got Accepted
INPUT IS ...
THERE ARE EMPTY LINES / LINES WITHOUT TWO FULL NAMES

I DON'T LIKE PROBLEMS LIKE THIS !!!!!!!!!!!!!!!!!!!!!!

i handled case second number in pair = 1 separetly.
it seems quite different in my solution.

Hint in 109 says you don't need to compute angles to sort by angles,
it is an optimization issue. I try to avoid angles.

i think you should subtract p2 from both p1 and p3. this should work when p1 p2 p3 counter-clockwise oriented. x1 -= x2; y1 -= y2; x3 -= x2; y3 -= y2; double angle1 = atan2(y1, x1); double angle2 = atan2(y3, x3); if( angle1 < 0.0 ) angle1 += 2*M_PI; if( angle2 < 0.0 ) angle2 += 2*M_PI; return 180.0*...

do you really need this angle ?
i did quite a lot of geometry and inconsistency was never a problem.
Check a hint for problem 109, bottom of problem description,
- detection of orientation.

i checked this 1 cousin removed 1 vs. 3 cousin case
and it's ok, but i get WA.
what doi printf for graph as in example and query:
a f

ok, i checked with sigsegv method and there are no such cases.

i'm confused, help pls.
what should i printf for this case:
d a
e a
d c
c b
e b
no.child no.parent
d e

Your solution is really good. Myself, i learned from this problem that judge accepts inline asm in c++. I just used three loops for logarithms and "jo" instruction to detect overflows. I think under-linear complexity algorithm is possible if "easy & dense" subset of ugly numb...