That’s a nice and polite way to put it ;)

Well, I am not sure if I have well understood the problem, but below is the hasty solution I came up with, to demonstrate my feeling that indeed, this algorithm is not fair as far as each position of the final rankings is concerned.

team = []
number_of_teams = 5
number_of_students = 150

for i in range(number_of_teams):
team.append([])

# for each ballot, find the favourable places
for ballot in range(15, 21):
n = 1
while (n <= number_of_students):
# how many students remain in this team
# that could belong to the first five
remaining_students = min(number_of_students – n + 1, number_of_teams);

for j in range(remaining_students):
# first student of the team
team[j].append(n+j)

# proceed to the next team
n += ballot

#for i in range(number_of_teams):
# team[i].sort()

print(team)

# initialize a double vector with zeros
# there has to be a better way
histogram = []
for i in range(number_of_teams):
histogram.append([])
for j in range(number_of_students):
histogram[i].append(0)

for i in range(number_of_teams):
for x in team[i]:
# start indexing from 1 (human readable)
# we already know that no student can achieve to be 0th in his year
histogram[i][x-1] += 1

print(histogram)

A 64-bit horse that can count
http://www.viva64.com/art-1-2-377673569.html

Thanks for yet another informative post. POSIX mandates a similar interface, ftw(3). But, to quote Issue 7, “The ftw() function may be removed in a future version.”. On the other hand it doesn’t say whether it will be replaced by some other interface.

Anyway, ftw(3) is implemented on top of fts(3) in NetBSD that I checked.

Cheers,
Stathis