It is time to rant some! The internet doesn’t contain enough ranting!

https://x.com/HandmaidsOnHulu/status/1889721210008654263?t=5yaLLgU-8jNlj8a4cBpZbQ&s=19

Hulu, April 8. So that’s correct in USA? What about the rest of the world?

Arthur C. Clarke: “The Wall of Darkness”.

Brave men do great deeds. Some deeds are dangerous. Some deeds are expensive.

Shervane, the protagonist, apparently has no women in his life. No wife and no children. He also apparently has nobody to say, maybe this risk and that expense should be avoided. To me this leaves a gaping hole in the story.

🇩🇰🇩🇰🇩🇰

Køn tildelt ved fødsel – når sproget forvansker virkeligheden

“Kønsdebatten er et minefelt, hvor man hurtigt kan detonere en lavine af karakterbeskyldninger ved ytringen af bestemte ord eller formuleringer.”

Ja, tænk. Ord betyder noget. Ords betydning ændrer sig. Visse ord hænger sammen med visse (forældede) holdninger og volder smerte, når de bliver brugt. Det nyeste her er sådan set, at det er blevet lettere at høre, når der bliver sagt av. “Gider du lige lade være med at bruge det der ord?”

Ja, der er nok nogen, der overreagerer. Sådan er det jo ofte. Ja, der er nok også nogen, der reagerer meget voldsomt den 1000. gang, de oplever noget smertefuldt. Også selvom det den her gang var en ærlig fejltagelse, en uskyldig smutter.

Ja, jeg taler her om folk, der i stigende grad siger: “Jeg er en kvinde, og det gør enormt ondt, når du kalder mig en mand.” Eller omvendt.

Hvis vi ellers antager, at du er et kærligt og hensynsfuldt menneske, så er du vel også interesseret i ikke at forvolde smerte?

“Man tvinges til at navigere i et sprog, der synes forvirrende og ulogisk, og hvor man hele tiden skal tænke omvendt for at forstå, hvad der faktisk tales om.”

Bare rolig. Det skal nok falde på plads, når du har lært det. Sådan er det jo med alt nyt.

“Op er faktisk ned. Ned er faktisk op. Transkvinder er faktisk mænd. Transmænd er faktisk kvinder.”

Ja, ja, der er den. En person siger: “Jeg er en (trans)kvinde.” Men du, der ikke har prøvet at være i hendes hoved, krop eller sko, ved bedre. Fordi den gamle definition på kvinde er den rigtige og bør ikke ændres. Hm.

“Udskiftningen af ord og betydninger er dog et virkelig smart, strategisk greb, hvis man gerne vil ophæve anerkendelsen af de to biologiske køn.”

Måske er der en strategi her. Måske er der bare: “Jeg er en kvinde. Vil du være sød at bruge det ord?”

Så har jeg også lyst til at spørge, hvordan de biologiske køn defineres. Det er vist ikke så nemt.

“Disse to sætninger beskriver den samme situation, men de opleves forskelligt:

• Den lyshårede transpige gik med ind i pigernes omklædningsrum
• Den lyshårede transidentificerende dreng gik med ind i pigernes omklædningsrum“

Igen. Ja, det tager lidt tid at lære det her. Men det skal nok komme.

“Kommunale familievejledere må altså forstå, at en borgers subjektive køns-selvopfattelse har fortrinsret over personens objektive, biologiske køn, og at de skal korrigere deres sprog herefter.”

Tja. Hvis jeg skifter navn, adresse, religion eller job, så forventer jeg, at du følger med. At jeg fx skiftede navn bygger på noget subjektivt. Det der navn føles mere som mig. Senere får det nogle bureaukratiske følger, fordi vi hertillands registrerer, ja, faktisk godkender folks navne. Men når folkeregistret har registreret det, så er det ligesom mere rigtigt. Objektivt måske. I det nævnte tilfælde ville folkeregistret have påpeget, at der var tale om en kvinde. Er det objektivt nok?

“Kvinder er de feminine. Denne mand er feminin. Ergo er manden en kvinde.”

Åh mand. Eller kvinde.

Kvinder er de feminine = kvinder er kvindelige = kvinder er kvinder. Tautologisk.

Denne mand er feminin = denne mand er kvindelig = denne mand er en kvinde. Sludder.

“Udbredelsen af kønsidentitetsideologien påvirker hele vores samfund: Når sproget kontrolleres, bliver alle berørt.”

Ordet kønsidentitetsideologien er interessant. En ideologi er sådan noget med, hvor meget hvem skal betale i skat, eller om der skal være et privat sundhedsvæsen. Det her spørgsmål er ikke ideologisk. Det har med eksistens at gøre. Ønsker vi, at samfundet, lovgivningen, og ja, sproget, lader transkønnede eksistere, sådan som de er?

Eller er det fordi, fascisme også kan være en ideologi? Seriøst, at religion/afstamning kan føre til dehumanisering af mennesker, og i sidste ende død, det er bare ideologi?

Æhm. Hele dit indlæg der er da et forsøg på at kontrollere sproget. Ikke?

Jeg bemærker i øvrigt, at indlægget i høj grad taler om transkvinder, der “i virkeligheden” er voldtægtsmænd, mandlige mordere osv. Der er ikke så meget plads til fredelige, lovlydige transkvinder. Og endnu mindre til fredelige, lovlydige transmænd. Hm.

This week the question is: Can You Bounce a Ball Like an AI?

… suppose you have a unit square that’s rotating about its center at a constant (nonzero) angular speed, and there’s a moving ball inside. The ball has a constant (nonzero) linear speed, and there’s no friction or gravity. When the ball hits an edge of the square, it simply reflects as though the square is momentarily stationary during the briefest of moments they’re in contact. Also, the ball is not allowed to hit a corner of the square—it would get jammed in that corner, a situation we prefer to avoid.

Suppose the ball travels on a periodic (i.e., repeating) path, and that it only ever makes contact with a single point on the unit square. What is the shortest distance the ball could travel in one complete loop of this path?

And for extra credit:

Again, you have a rotating unit square and bouncing ball inside.

By now, you’ve hopefully found the shortest repeating path for which the ball makes contact with a single point on the square. Let’s call this path length L₁.

The next shortest repeating path for which the ball makes contact with a single point on the square has length L₂. To be clear, L₂ > L₁.

What is the length L₂?

Highlight to reveal (possibly incorrect) solution:

Spreadsheet.

A few tests lead me to believe, the only valid solutions are loops coinciding with regular polygons within a unit circle. The point on the square will be in the middle of one of the sides.

The first polygon is simply the loop of going 2 in one direction and 2 back. L₁ = 4.

The next polygon is a triangle. Each side of this triangle is also a long side of a triangle, where the centre of the unit circle is the 3rd vertex. 2 sides have the length 1. The angle at the centre is 360°/3 = 120°. The angles at the other vertices are (180°-120°)/2 = 30°. The length of the unknown side, x, satifies x/sin(120°) = 1/sin(30°) <=> x = √3. The perimeter of the triangle is 3√3, about 5.196.

The spreadsheet continues this trend. The perimeter of the polygons grow towards 2π.

Oh wait. While a unit circle has radius 1, diameter 2, a unit square simply has width 1, right? Sigh. — All results above should be divided by 2.

This week the question is: Can You Spin the Graph?

You’re taking a math exam, and you’ve been asked to draw the graph of a function. That is, your graph must pass the vertical line test, so that no vertical line intersects your function’s graph more than once.

You decide you’re going to graph the absolute value function, y = |x|, and ace the test.

There’s just one problem. You are dealing with a bout of dizziness, and can’t quite make out the x– and y-axes on the exam in front of you. As a result, your function will be rotated about the origin by a random angle that’s uniformly chosen between 0 and 360 degrees.

What is the probability that the resulting graph you produce is in fact a function (i.e., y is a function of x)?

And for extra credit:

In a more advanced course, you’ve been asked to draw a 3D sketch of the function z = |x| + |y|. As you’re about to do this, you are struck by another bout of dizziness, and your resulting graph is randomly rotated in 3D space.

More specifically, your graph has the correct origin. But the true z-axis is equally likely to point from the origin to any point on the surface of the unit sphere. (Meanwhile, the x-axis is equally likely to point in any direction perpendicular to the z-axis. From there, the y-axis is uniquely determined.)

What is the probability that the resulting graph you produce is in fact a function (i.e., z is a function of x and y)?

Highlight to reveal (possibly incorrect) solution:

Basically we’re looking at a triangle (2 of the sides are y=x and y=-x), and we’re rotating the triangle around the origin. If the triangle rotates enough for 1 side to be below the x axis and 1 side above, we’re out. If we rotate enough, so that both sides are now below the x axis, we’re back in business. The good rotations are in degrees 0-45 , 135-225 and 315-360. This represents 45, 90 and 45 degree slices of the 360 degrees a rotation can be, or in other words, 180/360 = 0.5. This is also our probability of not breaking the function. The answer is 0.5.

And for extra credit:

Desmos. Figure 1, 2, 3, 4, 5, 6, 7.

This time it’s a pyramid shape, that has to stay all above (or all below) the x y plane. The figures show, how I modeled this pyramid in Desmos, letting it intersect with a unit circle. The tip of the z axis can roam, but once it moves outside the 4 sided figure with curvy edges (let’s call it Curvy), the function breaks. So the question is: How big is that figure?

In figure 3 I see a Curvy, bounded by 4 great circles. Letting the area of the triangle be a and the area of Curvy be b, we see that 4a + 3b = 1/2 sphere area. (On the other side of the sphere, we have the same picture and the same equation.) Further, the middle band, bounded by 2 great circles, consists of 2a + b = m part of sphere area, and the upper band consists of a + b = n part of sphere, where m + 2n = 1/2. m depends on the angle between the 2 great circles. Their planes are defined by z = x + y and z = – x – y. This can be computed via the angle k, cos(k) = 1/3. As seen in figure 4, this means k is about 70.5 degrees. So our 2 equations become approximately 2a + b = 70/360 and a + b = 55/360. This again means b is about 0.108 part of sphere.

There are 2 Curvys, just like there were 2 good areas in the first puzzle. So our answer ends up being 2 x 0.108 = 0.216. Well, doing it with more digits, closer to 0.216347.

I am not at all sure about this one. It was hard for me.

This week the question is: Can You Hop to the Lily Pad?

You are a frog in a pond with four lily pads, marked “1” through “4.” You are currently on pad 2, and your goal is to make it to pad 1. From any given pad, there are specific probabilities that you’ll jump to another pad:

• Once on pad 1, you will happily stay there forever.
• From pad 2, there’s a 1-in-2 chance you’ll hop to pad 1, and a 1-in-2 chance you’ll hop to pad 3.
• From pad 3, there’s a 1-in-3 chance you’ll hop to pad 2, and a 2-in-3 chance you’ll hop to pad 4.
• Once on pad 4, you will unhappily stay there forever.

Given that you are starting on pad 2, what is the probability that you will ultimately make it to pad 1 (as opposed to pad 4)?

And for extra credit:

Once again, you are a frog in a pond. But this time, the pond has infinitely many lily pads, which are numbered “1,” “2,” “3,” etc. As before, you are currently on pad 2, and your goal is to make it to pad 1, which you would happily stay on forever.

Whenever you are on pad k, you will hop to pad k−1 with probability 1/k, and you will hop to pad k+1 with probability (k−1)/k.

Now, what is the probability that you will ultimately make it to pad 1?

Highlight to reveal (possibly incorrect) solution:

Program.

This problem can also be expressed as:

The initial state is a = (0 1 0 0). We are on the 2nd pad with probability 100%.

Then a matrix describes each state change:

    ⌈  1   0   0   0  ⌉
M = | 1/2  0  1/2  0  |
    |  0  1/3  0  2/3 |
    ⌊  0   0   0   1  ⌋

After 1 jump, the state is a * M = (1/2 0 1/2 0). What we’re looking for is the state after infinitely many jumps. It’s actually okay to simulate 200 jumps, this converges nicely. The value of a * M²⁰⁰ = (0.6 0 0 0.4). The value we’re looking for is 0.6. This is reflected in the first line of output from my program.

And for extra credit:

Repeat the above with ever bigger ponds. I stopped at 20 pads. My program output shows, that the value we’re looking for converges. The last line says 0.63212055882856. I’m not sure what that signifies. e/pi²?

This December I participated in Advent of Code.

So. Looking back. Days 1-10 were fun, because it was fun to code again. Day 16 (1 point for walking forwards, 1000 to turn) was fun, after the fact, because I found an elegant way (with a nice diagram) to represent the labyrinth as a graph, and finally just implemented Dijkstra fully. Day 17 (build a small programming language) was very satisfying to solve, even if it involved a lot of paper. For day 19 (towels) I had to learn memoization. (I never learned this before?) Day 20 (a labyrinth that turns out just to be one long path, and it’s possible to walk through walls) it was an epiphany, that (part 1 of route) + (part 2 of route) – (discarded part in the middle) + (shortcut) was a way to solve the puzzle. Day 24 (a lot of gates) required some paperwork.

This is my attempt to fill out the complete personal times for me. I’m a little sad, that it simply says “more than 24 hours” officially.

Part 1 of days 17, 22 and 23 must have been easy, I spent minutes getting them right (just 2 or 7 or 8 days late). As I remember it, day 25 didn’t take a long time either, but I wanted to be well rested before I attempted it, so, happy surprise.

      --------Part 1--------   --------Part 2--------
Day       Time   Rank  Score       Time   Rank  Score
 25 8:03:01     22113      0 8:03:02     14148      0
 24 7:10:58     23368      0 8:11:00     15862      0
 23 7:20:30     23812      0 8:10:15     21472      0
 22 8:17:59     25204      0 8:20:00     21799      0
 21 9:10:07     19594      0 9:17:22     16355      0
 20 5:05:00     24815      0 6:02:28     21577      0
 19 4:06:00     28005      0 4:08:12     25333      0
 18 4:11:11     27991      0 4:11:49     27338      0
 17 2:09:34     28700      0 3:08:04     20591      0
 16 2:06:05     24090      0 3:08:50     25303      0
 15   01:19:07   5843      0   07:18:20   8861      0
 14   01:28:12   7258      0   02:13:42   6000      0
 13   03:04:50  11545      0   03:08:52   7714      0
 12   03:55:52  14298      0   04:04:26   7259      0
 11   00:24:58   6479      0   01:09:10   5330      0
 10   00:44:34   6515      0   00:48:19   6013      0
  9   01:26:25   9059      0   02:59:32   8039      0
  8   01:53:25  10087      0   02:04:40   9294      0
  7   17:55:39  47349      0   18:02:03  44377      0
  6   00:45:55   8375      0   02:18:28   8078      0
  5   13:07:46  55110      0   13:13:45  44839      0
  4   01:51:36  15622      0   02:06:52  13570      0
  3   16:51:27  91593      0   17:04:46  80436      0
  2   12:11:26  78108      0   13:30:49  60978      0
  1   17:12:43  92846      0   17:24:33  87414      0

This is, for completeness, based on these timestamps:

-rw-rw-r-- 1 slimbook slimbook      6773 Dec 19 14:50 d16c.php
-rw-rw-r-- 1 slimbook slimbook      4921 Dec 19 15:34 d17a.php
-rw-rw-r-- 1 slimbook slimbook     93787 Dec 20 14:04 day17b.ods
-rw-rw-r-- 1 slimbook slimbook      6166 Dec 22 17:11 d18a.php
-rw-rw-r-- 1 slimbook slimbook      6679 Dec 22 17:49 d18b.php
-rw-rw-r-- 1 slimbook slimbook      2092 Dec 23 08:43 d19a.php not right!!!
                                                12:00 ?
-rw-rw-r-- 1 slimbook slimbook      1075 Dec 23 16:12 d19b.php
-rw-rw-r-- 1 slimbook slimbook      4482 Dec 25 11:00 d20b.php
-rw-rw-r-- 1 slimbook slimbook      4393 Dec 26 08:28 d20c.php
-rw-rw-r-- 1 slimbook slimbook      9763 Dec 30 16:07 d21c.php
-rw-rw-r-- 1 slimbook slimbook     10164 Dec 30 23:22 d21e.php
-rw-rw-r-- 1 slimbook slimbook      1371 Dec 30 23:59 d22a.php
-rw-rw-r-- 1 slimbook slimbook      1951 Dec 31 02:00 d22b.php
-rw-rw-r-- 1 slimbook slimbook      1645 Dec 31 02:30 d23a.php
-rw-rw-r-- 1 slimbook slimbook       783 Dec 31 16:15 d23tmp.txt

-rw-rw-r-- 1 slimbook slimbook      1878 Dec 31 16:58 d24a.php
-rw-rw-r-- 1 slimbook slimbook      7940 Jan  1 17:00 d24mermaid3.txt
-rw-rw-r-- 1 slimbook slimbook      1217 Jan  2 09:01 d25a.php
                                                09:02

This December I participated in Advent of Code.

Part 1 was sort of easy. Check whether a lock pin and key pin overlap in this particular combination, and go through all the possible combinations, counting. Part 2 turned out to be: Complete all the 49 other parts this December, and click this link!

Part 1:

<?php

///////////////////////////////////////////////////

$input = file_get_contents('./d25input2.txt', true);

$item = 0;
foreach(preg_split("/((\r?\n)|(\r\n?))/", $input) as $line) {
  if(strlen($line)>2) {
    $items[$item][] = str_split($line);
  } else {
    $item++;
  }
}

foreach($items as $item) {
  $tmp = implode($item[0]);
  if($tmp == "#####") {
    $locks[] = $item;
  } else {
    $keys[] = $item;
  }
}

foreach($locks as $id => $lock) {
  // each pin, 5 of them
  for($i=0;$i<5;$i++) {
    // each pin element, 7 of them
    $pinh = 0;
    for($j=1;$j<7;$j++) {
      if($lock[$j][$i] == "#") {
        $pinh++;
      }
    }
    $locksig[$id][] = $pinh;
  }
}

foreach($keys as $id => $key) {
  // each pin, 5 of them
  for($i=0;$i<5;$i++) {
    // each pin element, 7 of them
    $pinh = 0;
    for($j=5;$j>0;$j--) {
      if($key[$j][$i] == "#") {
        $pinh++;
      }
    }
    $keysig[$id][] = $pinh;
  }
}

$lockkeymatch = 0;

foreach($locksig as $lock) {
  foreach($keysig as $key) {
    $tmp = array();
    for($i=0;$i<5;$i++) {
      $tmp[] = $lock[$i] + $key[$i];
    }
    if(max($tmp) <= 5) {
      $lockkeymatch++;
    }
  }
}

print("$lockkeymatch\n");

?>

A new December and a new bunch of puzzles from mscroggs.co.uk.

All the hints:

And I know no. 8 and 23 are wrong.

My first look at this problem:

• (aaa – bbb)*c = x
• (x – ddd)*e = y
• (y – fff)*g = hhhhh

Let’s sort this a little bit.

Very clear, and as I suspected, very late in the game. aaa = 444.

Or to put it another way:

• We’re looking for bbb, ddd and fff.
• One of these is 144.
• One of these is 138. (6 * 23)
• One of these is a multiple of 990, therefore one of these: 110, 165, 198, 330, 495, 990. There’s no overlap with the 2 we already have, so this must be the third one.
• One of the digits of ddd is 9. This must be the “multiple of 990” one.
• As 256 is 2⁸, ddd must be odd.
• Combining these 3 facts, we get ddd = 495.
• As 851 = 23 * 37, fff can’t be 138. fff = 144.
• bbb = 138.

c can’t be 2, 7, 5, 3, 9 or 6. This leaves 1, 4 and 8. (I assume for now, that none of the 1 digit numbers are 0.)

e can’t be 1, 7, 9, 4, 5, 2, 8, 6. This leaves 3.

g can’t be 9, 1, 2, 7 or 4. This leaves 3, 5, 6 and 8.

Let’s repeat that:

• (aaa – bbb)*c = x
• (x – ddd)*e = y
• (y – fff)*g = hhhhh
• aaa = 444
• bbb = 138
• c is 1, 4 or 8
• ddd = 495
• e = 3
• fff = 144
• g is 3, 5, 6 or 8

Now it’s just a question of going through all the options. Discarding the negative options for hhhhh first, and testing the true 5 digit options next.

Hey! Turns out, (((444-138)*4-495)*3-144)*8 = 16344.

This December I participated in Advent of Code.

One of those days, where I probably grumbled a lot. Part 1 was easy. I chose an algorithm, where I went through a list of all the expressions, checked whether the gate’s output could be calculated (and if yes, deleted the expression from the list), looped back and did it again, until the list was empty. Part 2… I tried several different approaches. Reading hints from others, I ultimately used a free online diagram tool, Mermaid, that accepts text input. Perfect. I wrote a small program to turn the input into something Mermaid could understand (every value, value, gate, value, turned into 2 x value, arrow, value). Then I tried to analyze this diagram. Printed a lot of full adders. And began copying the diagram on to these. Every time the diagram wouldn’t fit, I’d found a swap. Then it took some thinking to figure out how to swap back. But I got there!

Part 1:

<?php

///////////////////////////////////////

$input = file_get_contents('./d24input2.txt', true);

foreach(preg_split("/((\r?\n)|(\r\n?))/", $input) as $line) {
  if(strlen($line)>2) {
    // x00: 1
    if(preg_match("/^(\w+): (\d)$/", $line, $m)) {
      $values[$m[1]] = $m[2];
    }
    // ntg XOR fgs -> mjb
    if(preg_match("/->/", $line, $m)) {
      $m = explode(" ", $line);
      $expr[] = $m;
    }
  }
}

while(sizeof($expr) > 0) {
  foreach($expr as $key => $exp) {
    $val1 = $exp[0];
    $val2 = $exp[2];
    if(isset($values[$val1]) && isset($values[$val2])) {
      $gate = $exp[1];
      $val3 = $exp[4];
      switch($gate) {
        case "AND":
          $values[$val3] =
            $values[$val1] * $values[$val2];
          break;
        case "OR":
          $values[$val3] =
            $values[$val1] + $values[$val2] 
            - $values[$val1] * $values[$val2];
          break;
        case "XOR":
          $values[$val3] =
            ($values[$val1] + $values[$val2]) % 2;
          break;
      }
      unset($expr[$key]);
    }
  }
}

$allzs = $values;
foreach($allzs as $key => $posz) {
  if(!preg_match("/z/", $key, $m)) {
    unset($allzs[$key]);
  }
}

print_r($allzs);

$sum = 0;
foreach($allzs as $key => $z) {
  $pos = (int) str_replace("z", "", $key);
  $sum += $z * pow(2, $pos);
}

print("$sum\n");
?>

Part 2:

Big diagram. Small diagrams.

A new December and a new bunch of puzzles from mscroggs.co.uk.

All the possible 3 digit numbers are here. E.g. we have 111, 211, 311, 411, 511, 611 and 711. For these the mean of the first digit is 4. Expanding this to all the possibilities, the mean for every digit is 4. The result is 444.

This December I participated in Advent of Code.

For part 1, I optimized a little. Like only storing the representation of each 3 way computer group once. Also 3d arrays. For part 2… I flirted a little with BK (ehm, can’t remember what that stands for), but that program seemed to never stop. So I did it differently. Just like part 1 asked me to create all the 3 way groups, I now moved on to create all the 4 way groups, 5 way groups etc. As each computer had about 13 friends, this shouldn’t take forever to get through. I think it took 35 minutes. I stopped when I only had 1 group of the current size. Oh yeah, and I changed the data structure a little. (BK was sort of the other way around. Trying to grow 1 large set greedily.)

Part 1:

<?php

//////////////////////////////////

$input = file_get_contents('./d23input2.txt', true);

foreach(preg_split("/((\r?\n)|(\r\n?))/", $input) as $line) {
  if(strlen($line)>2) {
    $conn1[] = explode("-", $line);
  }
}

foreach($conn1 as $comp) {
  $comp1 = $comp[0];
  $comp2 = $comp[1];
  $conn2[$comp1][$comp2] = 1;
  $conn2[$comp2][$comp1] = 1;
}

foreach($conn2 as $comp1 => $c1con) {
  foreach($c1con as $comp2 => $cona) {
    foreach($c1con as $comp3 => $conb) {
      // make sure comp2 and comp3 are different
      if(isset($conn2[$comp2][$comp3])) {
        // only store the 3 way connection alphabetically, so as not to store it 6 times
        if(strcmp($comp1,$comp2) < 0 && strcmp($comp2,$comp3) < 0) {
          // only store this set, if one of the computers begin with t
          if(substr($comp1, 0, 1) == "t" || substr($comp2, 0, 1) == "t" || substr($comp3, 0, 1) == "t") {
            $conn3[$comp1][$comp2][$comp3] = 1;
          }
        }
      }
    }
  }
}

$sum = 0;
foreach($conn3 as $a) {
  foreach($a as $b) {
    foreach($b as $c) {
      $sum += $c;
    }
  }
}
print("$sum\n");

?>

Part 2:

<?php

//////////////////////////////////

$input = file_get_contents('./d23input1.txt', true);

foreach(preg_split("/((\r?\n)|(\r\n?))/", $input) as $line) {
  if(strlen($line)>2) {
    $conn1[] = explode("-", $line);
  }
}

foreach($conn1 as $comp) {
  $comp1 = $comp[0];
  $comp2 = $comp[1];
  $conn2[$comp1][$comp2] = 1;
  $conn2[$comp2][$comp1] = 1;
}

foreach($conn2 as $comp1 => $c1con) {
  foreach($c1con as $comp2 => $cona) {
    foreach($c1con as $comp3 => $conb) {
      // make sure comp2 and comp3 are different
      if(isset($conn2[$comp2][$comp3])) {
        // only store the 3 way connection alphabetically, so as not to store it 6 times
        if(strcmp($comp1,$comp2) < 0 && strcmp($comp2,$comp3) < 0) {
          //$conn3[$comp1][$comp2][$comp3] = 1;
          $conn3b[] = array($comp1, $comp2, $comp3);
        }
      }
    }
  }
}

$allcomp = array_keys($conn2);

$multiple = sizeof($conn3b);
$connx = $conn3b;
while($multiple > 1) {
  print("No. of sets in the iteration: $multiple\n");
  $conny = array();
  // build new arrays of length y from length x - y = x+1
  foreach($connx as $wayx) {
    foreach($allcomp as $key => $newcomp) {
      // test that computer is not already in array
      if(!in_array($newcomp, $wayx)) {
        $newgood = 1;
        // test that new is connected to all in wayx
        foreach($wayx as $oldcomp) {
          if(!isset($conn2[$oldcomp][$newcomp])) {
            $newgood = 0;
            break;
          }
        }
        if($newgood == 1) {
          // create new wayy
          $wayytmp = $wayx;
          array_push($wayytmp, $newcomp);
          sort($wayytmp);
          if(!in_array($wayytmp, $conny)) {
            $conny[] = $wayytmp;
          }
        }
      }
    }
  }
  
  $multiple = sizeof($conny);
  if($multiple == 1) {
    print("Max set found!\n");
    print_r($conny);
  }
  $connx = $conny;
}

?>