TL;DR

Here we are with TASK #1 from the Perl Weekly Challenge #104. Enjoy!

The challenge

Write a script to generate first 50 members of FUSC Sequence. Please refer to OEIS for more information.

The sequence defined as below:

fusc(0) = 0
fusc(1) = 1
for n > 1:
when n is even: fusc(n) = fusc(n / 2),
when n is odd: fusc(n) = fusc((n-1)/2) + fusc((n+1)/2)

The questions

My main questions here are regarding the inputs and the outputs, so they’re quite boring:

  • is the input bound to be a non-negative integer?
  • what is the expected output? Something to be printed or just an array?

The solution

We’ll take a look at two solutions, just for fun. Wheeee!

The first one is quite adherent to the definition above: calculate each new item based on the recursive formula. This does not need to be implemented recursively though, a simple iteration will do:

sub fusc ($n) {
   return [0 .. $n - 1] if $n <= 2;
   my @fusc = (0, 1); # fusc(0), fusc(1)
   while (@fusc < $n) {
      push @fusc, $fusc[@fusc >> 1];
      last if @fusc >= $n;
      push @fusc, $fusc[-1] + $fusc[1 + @fusc >> 1];
   }
   return \@fusc;
}

There’s a bit to unpack here, because it’s admittedly not in the best interest of the reader (I just felt like this):

  • there’s no explicit index variable for tracking the item we have to add. Adding an item in a Perl array can be done with push (so we don’t need it) and its length can be taken by evaluating it in scalar context. Who needs those stinking, readable variables then?!?
  • Dividing by 2 in an integer way is the same as a bit shift left of one position, which accounts for @fusc >> 1. This expression also provides the scalar context we need to get the number of elements in @fusc, yay!
  • There are two ways to exit from the loop, one for even input $n and another one for odd values. Try to figure out which is which.
  • For odd positions, the value in position $(n - 1) / 2$ is (by definition!) the same as the value in position $(n - 1)$, which in Perl terms translates into $fusc[-1]

As a twist, I thought… why not go the other way around? I mean, all (well, most) elements in the sequence directly affect three other elements. In other terms, for $k$ sufficiently large, the $k$-th element affects elements at positions $(2 k - 1)$, $2k$, and $(2k + 1)$.

There is an important caveat here, because this does not apply for low values of $k$. Actually, fusc(1) does indeed affect fusc(2) and fusc(3)… but it MUST NOT not affect fusc(1) 🙄

All of this allows us to build a constructive solution from the ground up:

sub fusc_sieve ($n) {
   my @fusc = (0, 1);
   for my $i (1 .. $n >> 1) {
      $fusc[$i * 2]      = $fusc[$i];
      $fusc[$i * 2 + 1] += $fusc[$i];
      $fusc[$i * 2 - 1] += $fusc[$i];
   }
   $fusc[1] = 1;
   return [@fusc[0 .. $n - 1]];
}

Again, we’re using $n >> 1 instead of int($n / 2) just to be a bit snob and stimulate that WTF?!? feeling in the reader.

The 1-exception is taken care of in two places here:

  • the most obvious one is when we (re)set the value of fusc(1) explicitly just after the loop;
  • the hidden one lies in the order used to updated the three affected values. Note that the value $fusc[$i * 2 - 1] is updated last… this allows us to defer the vicious self-updating of the element 1 last, so that this glitch does not affect any other element down the line.

Our last attention point is on the returned value: this method tends to generate elements a bit generously, so we trim the output a bit when we return it.

All in all, taking this route was a fun sub-challenge, although a perilous one. Which is a lesson in itself: simpler solutions are often better for a reason.

Stay safe everyone!