Adam Leadbetter

Uses function isleapyear on lines 61 & 64. However, that function does not appear to be a matlab function.

Fixed the problems I saw. It does what it claims to do. It is faster now too, taking a much smaller chunk of time away from your own computations. Whereas the last time I did this test, it too almost 10 seconds just to do the progress report, this test is down to only about a second.

tic,for i = 1:10000,perccount(i,10000),end,toc
Elapsed time is 1.243821 seconds.

Perccount is not infinitely fast, but waitbar will be just as slow, if not worse. This begs the question of just how slow?

h = waitbar(0,'I''m busy!');
tic
n = 10000;
for i = 1:n
 waitbar(i/n)
end
delete(h)
toc
Elapsed time is 124.846332 seconds.

Yes, just 10000 calls to waitbar, with nothing else in the loop, took over 2 minutes!

So, for the person who will have wanted to ask, "Why not just use waitbar?" this is a good reason. All of those calls to update a slider bar take considerable time. Yes, I'll admit that more careful coding, calling waitbar only infrequently in that loop would have been faster. So,

h = waitbar(0,'I''m busy!');
tic
n = 10000;
for i = 1:n
  if (rem(i,100) == 0)
    waitbar(i/n)
  end
end
delete(h)
toc

Elapsed time is 1.280422 seconds.

So for the person who does not wish to write the extra code to make waitbar function passably, or for the person who would actually prefer to have progress reported in the command line, perccount is a very viable option.

My thanks for fixing this. Well done.