explain xkcd - User contributions [en]

Talk:3122: Bad Map Projection: Interrupted Spheres

2025-09-01T07:57:33Z

88.129.22.189:

Where's Greenland? [[User:SubtrEM|SubtrEM]] ([[User talk:SubtrEM|talk]]) 20:00, 30 July 2025 (UTC)
:It is on the backside of North America globe of course. --[[User:Kynde|Kynde]] ([[User talk:Kynde|talk]]) 07:49, 31 July 2025 (UTC)

My first edit! Hope it's ok! [[User:Jkusa.jr| Jkusa.jr]] 08:12, 30 July 2025 (UTC)

Wait, maybe I'm crazy, but I feel like there's actually a good map idea here? If you made proper globes centered on each continent.... Does this exist? [[Special:Contributions/2601:241:8002:3E0:CE5:D9D:CF64:76FD|2601:241:8002:3E0:CE5:D9D:CF64:76FD]] 21:36, 30 July 2025 (UTC)
:Well... that is what a globe is. You just turn it until it is on the continent you wish and look from the right angel ;-) Drawing a globe on a paper does nothing to remove the distortion from normal flat maps. --[[User:Kynde|Kynde]] ([[User talk:Kynde|talk]]) 07:49, 31 July 2025 (UTC)
:Google "globus polski", somebody already came up with this kind of joke in the communist Poland [[Special:Contributions/2A00:F41:80A5:A62:0:59:BA67:7B01|2A00:F41:80A5:A62:0:59:BA67:7B01]] 02:43, 4 August 2025 (UTC)

Not sure what the joke is! Another dis of the concept of continents? [[Special:Contributions/2A09:BAC3:9C1B:1955:0:0:286:B1|2A09:BAC3:9C1B:1955:0:0:286:B1]] 22:01, 30 July 2025 (UTC)
:It is a bad map projection. That is a joke in it self. Another way to badly draw a map, that is the ongoing joke. Of course there is also the silly joke in the title text like the Earth is actually spread over 7 spheres. --[[User:Kynde|Kynde]] ([[User talk:Kynde|talk]]) 07:50, 31 July 2025 (UTC)
::The thing is, attempts to translate a full sphere to a flat map requires various compromises (e.g. points that are close in real life may appear far apart on the map). Translating subsets of a full sphere out onto their own complete spheres technically requires the ''opposite'' type of compromises (e.g. points that are far apart in real life are now very close...). And add to that inevitable choice of angular and area distortions. [[Special:Contributions/82.132.244.235|82.132.244.235]] 10:12, 31 July 2025 (UTC)
:::It sometimes ''does'' feel as if a lot of people around the world were living on different planets, so we might as well go all out with the concept. Makes sense to me. [[User:PaulEberhardt|PaulEberhardt]] ([[User talk:PaulEberhardt|talk]]) 11:59, 31 July 2025 (UTC)

Yet another map that ignores/erases New Zeeland. {{unsigned|ProfKrueger|01:28, 31 July 2025}}
:No it doesn't ignore it. Just like Greenland is on the other side of the globe in North America so is NZ on the other side of Australias globe. You cannot either see Sydney or Tasmania or LA or other Western states in the US. In Europe you cannot see most of Scandinavia (although the important part is there ... Denmark ;-) and Madagascar is also left our near Africa as is the entire middle east and most of Russia. --[[User:Kynde|Kynde]] ([[User talk:Kynde|talk]]) 07:49, 31 July 2025 (UTC)

Thought Randall was becoming political leaving out N.Korea, Middle east, and Russia; then noticed the colossal China...--[[User:Darth Vader|Darth Vader]] ([[User talk:Darth Vader|talk]]) 09:42, 31 July 2025 (UTC)
:He also shows, the soon to be dictator ruled country, US, so he is not having any problems showing those kind of countries :-p --[[User:Kynde|Kynde]] ([[User talk:Kynde|talk]]) 19:49, 31 July 2025 (UTC)

What are those islands that I see in the center of the Antarctica globe? [[Special:Contributions/67.82.132.47|67.82.132.47]] 13:15, 31 July 2025 (UTC)
:The largest is {{w|Berkner Island}}. Mostly trapped in the RF-Ice Shelf, under normal viewing conditions and even many maps, but here represented with the whole of the Weddel Sea as if 'open' all the way through it, and not just featureless ice-sheet over both 'land' and watee. Over on the other side of Palmer's Land, as another notable feature (the characteristic long peninsula), is Alexander Island, and other semi-ilsand masses; up-of-centre on the left, as we see the Antarctic-Globe view. [[Special:Contributions/82.132.245.41|82.132.245.41]] 14:12, 31 July 2025 (UTC)

I love this! Bad Map Projection returns, now with all 7 continents on 7 globes. Still better than the Gall-Peters Projection! [[User:Strontium|Strontium]] ([[User talk:Strontium|talk]]) 19:45, 31 July 2025 (UTC)

It seems as if the continents aren't centered on their individual globes. Should they be? [[User:These Are Not The Comments You Are Looking For|These Are Not The Comments You Are Looking For]] ([[User talk:These Are Not The Comments You Are Looking For|talk]]) 01:57, 3 August 2025 (UTC)
:Can anything ever be centred on a globe? To my mind, though, the visible bit of the continent is (at least in the centre of the view, before the spherical distortion and wrapping over the 'horizon' of each) positioned as taken from a flat projection (no idea which, without drilling into centre-to-centre distances and angles). Perhaps Randall went for artfully-placed spheres, ''roughly'' aligned to his source flat-projection of choice, then 'wrapped them around from where the actual sphere-nearpoint landed.
:Also, estimations of where the flat-map centre of the continent lies can be difficult to agree. The exact mid-latitude and mid-longitide between its extremes, or a geometric 'weighted' average (of either the spherical-segment or flat map)? Is that of just the 'mainland' continent, or including all the way out to the remotest islands off the edge of an extended contintal shelf? Or just so that there is roughly equal expectations of land (major island chains counting more than minor peninsulae) hidden behind the western/northern limbs of the sphere as with the counterpart eastern/southern ones. Or just how it's placed so that it's a more aesthetic+recognisable view of the continent than if it ''was'' any given mathematically determined geographical centre at the face-on centre of the sphere? So many design/layout choices, and I'm not sure which one I'd choose if I came to doing it myself. I'd probably be more 'logical' than artistic, perhaps find the centre-of-gravity of the mass, but it might not look as good as just 'rolling the marbles around' until it looked neat. Or a bit of dry and precise analysis ''with'' artful adjustment, which is what Randall's style seems to be. [[Special:Contributions/82.132.237.7|82.132.237.7]] 15:51, 3 August 2025 (UTC)

What does "smartass response" mean? This is a confusing explanation. [[Special:Contributions/2603:800C:1200:596A:AB18:B71B:2BC5:5651|2603:800C:1200:596A:AB18:B71B:2BC5:5651]] 03:34, 6 August 2025 (UTC)
:As per the linked previous comic, "A Globe / Yes, you're very clever"... Not how I'd word it, but I can't see how it's not plain enough.
:Is this the sole point of your confusion, or is that a more general observation that you tacked on to your initial question? Help us to help you. [[Special:Contributions/82.132.247.119|82.132.247.119]] 13:09, 6 August 2025 (UTC)
:I made sure to replace it with "wisecrack" since the incomplete description notice called for more formal writing. [[Special:Contributions/2601:300:4083:1C70:F1B:D867:C7C7:F254|2601:300:4083:1C70:F1B:D867:C7C7:F254]] 06:43, 12 August 2025 (UTC)

The "landbridge" text leads to questions of America being one or two continents, or for that matter Europe, Asia, and Eurasia, as well as Asia/Africa.

Talk:3125: Snake-in-the-Box Problem

2025-09-01T07:46:50Z

88.129.22.189:

The math problem in question is https://oeis.org/A099155 [[User:Mei|Mei]] ([[User talk:Mei|talk]]) 21:57, 6 August 2025 (UTC)

why is d>8 unsolved? stevethenoob 21:59, 6 August 2025 (UTC)
:Computational power, I guess, although I'm going to go out on a limb and predict that for N=9 snake=196. [[Special:Contributions/94.73.52.245|94.73.52.245]] 23:18, 6 August 2025 (UTC)
:It's not that hard to imagine: if you were to try a brute force search it would take time that's exponential in the path length, which itself is exponential in d. There are evidently methods to do it slightly better, but not enough to make solving d=9 feasible yet. [[User:Zmatt|Zmatt]] ([[User talk:Zmatt|talk]]) 10:03, 7 August 2025 (UTC)
:To give an impression of the rate at which these get solved: d=6 was solved in 1988, d=7 in 1996, d=8 in 2014. [[User:Zmatt|Zmatt]] ([[User talk:Zmatt|talk]]) 10:32, 7 August 2025 (UTC)
:Computational power is right. Time complexity is *super*-exponential - even the number of *nodes* increases exponentially as increasing N by 1 doubles the nodes. And time complexity doesn't increase linearly with the number of nodes - if we imagine a brute force algorithm there's 2^n nodes and each node has n-1 options, so we're looking at a multiple of exponential time. Current lower bound for N=9 is 190. [[User:GorillaWarfare|GorillaWarfare]] ([[User talk:GorillaWarfare|talk]]) 06:57, 10 August 2025 (UTC)

I would argue that computer science has one as well with the China room problem. [[User:Ctinsman|Ctinsman]] ([[User talk:Ctinsman|talk]]) 22:14, 6 August 2025 (UTC)

: Humans aren't cute animals (mostly), so I propose a variant of the problem called the Chinese Red Panda Room [[Special:Contributions/177.12.49.23|177.12.49.23]] 22:38, 6 August 2025 (UTC)

Interesting. Just a few days ago I was investigating a very similar idea (looking at a path that transitioned between adjacent ''faces'' of a polyhedron, which was effectively going from vertex to connected vertex upon that chosen polyhedron's ''dual''), but for the opposite reason, i.e. looking for the paths that actually maximised proximity (along the path) between neighbouring faces (upon the polyhedra), so that it actually minimised the search back/forth along the path-chain to establish what value the adjacent polyhedron faces (beyond the ones automatically at ±1 positions on the chain) inherited.
 As to solving this one (basically disallowing visiting of any nodes adjacent to prior visits ''other'' than the single one that the +1 position of the chain has to first go to), I've got a basic idea of how I'd N-dimensionally space-search the possible routes (after all, visiting any given node at {0,1} value for dimensions [a, b, c, ...] rules out now visiting all of [!a, b, c, ...], [a, !b, c, ...], [a, b, !c, ...], etc, ''except'' whichever one of these was chosen for the next step of onward travel), for valid foldings across the appropriate N-polytype cuboidal analogue. Though I suspect that the exponental (or greater!) growth in the potential search-trees you'd use would be the sticking point. No point in setting off an exhaustive algorithm if it seemed likely to take three years to check just 1% of possibilities, and no doubt more dedicated analysis than my own brute-forcing method has already hit other problems in trying a more nuanced extrapolation between each level of added dimensionality, which is where the unsolved nature of this starts to bite.
 But also think it'd be far more interesting to investigate the possibilities in the N>3-Dimensional extensions of non-cubic platonic solids, like the {{w|600-cell}} and beyond, and establish what allowable lengths of traversal ''they'' would allow, under similar stipulations.
 Great! I love getting things like this to think about. If I can spare the time needed... [[Special:Contributions/82.132.245.59|82.132.245.59]] 22:22, 6 August 2025 (UTC)
: I think you've been nerd-sniped. [[Special:Contributions/177.12.49.23|177.12.49.23]] 22:42, 6 August 2025 (UTC)
::So far, I've personally got as far as:
::*For any given number of dimensions, N, there are always N adjacent points (point, zero dimensions, zero neighbours; line, one dimension, one neighbour; square, two dimension, two neighbours, etc).
::*In total, there are 2N points (0d=1, 1d=2, 2d=4, etc)
::*A maximum possible length, L, has a ''lower'' lower limit of starting at any particular vortex and only taking directions that are perpendicular to all prior directions (for a cube, only go by x, y and z directions once), and this would be eaual to N.
::*But that's overly-lazy, as you're ruling out (as you gain enough dimensions) revisiting a dimensional plane, even though you're allowed to revisit a point on that plane that's shifted by at least ''two'' other dimensions of offset. e.g. the top right of a cube's facing face when you started at the bottom left of it (went 'deep' to the rear face, took two steps from the rear-lower-left to rear-upper-right then back).
::*For the first step, you have N choices from your starting position. You take one and cannot later visit any of the ones you did not choose to go to. For the second step onwards, you have N-1 basic choices (every direction but backwards to the prior step) and should choose one and rule out ever visiting the rest.
::*This gives a new (at least for N>2) lower limit to L whereby the sum of starting, taken and not-taken nodes that you count can be added to by new steps until you would end up with have a total greater than 2N. (Line: start on one (of two), choice of one (taken), two points 'marked', only two points possible; Square: start on one (of four), two choices, take one, reserve one, three points 'marked', still the fourth point available for L=2, but then five points would be marked (the untaken-from-start being the only non-backwards choice) so can't go further.
::*But this is also wasteful as (in increasingly higher dimensions) there's nothing to stop an unvisited neighbour of a past step from being a(n enforced) unvisited neighbour from a later step, as you 'choose' to go only to a valid further point. So clever "near-neighbour" backtracking can reduce the number of ''freshly'' eaten-up points and thus maintain more future points for more steps.
::**Noting that past-step no-go-neighbours that can possibly 'fold into' the current-step's not-going-neighbours list only become such after ''at least'' two intervening steps (for 'square-based' hypercubic domains, whereas triangle-based hypernets (e.g. tetra-, octa- and icosohedrons, in 3-space) happen after just one step, and pentagon-based ones (dodecahedrons in 3-space) can't take advantage of this in less than three. (This seems to share some of the mathematics with the 'classical' rabbit-population problem, whereby new offspring only become viable breeding population after a step or two since their generation.)
::*Optimally, in fact, you should aim to double-back in such a way as leave yourself with ''all but one'' onward neighbour unavailable (thus only eating up potential points at a rate of one per step, at that point).
::*Heading vaguely back 'towards' past snake-lengths, in higher-dimensional hypernets, seems like the best(/longest) space-filling strategy. It's a bit like coil-built pottery, but with more undulations (and dimensions) to it. But with care to make sure you don't burrow yourself into a dead-end with ''no'' viable onward choices while still having maybe half of the potential visitable/neighbourable points untouched, or avoiding filling 'voids' to guaranteeing accessing a majority of the potential future visits, but unwisely not exploiting all the phase-space of vertices optimally.
::**I can mentally visualise doing this successfully in 3-, 4- and 5-cube situations, elegantly enough (it's like , but N>=6 versions get increasingly hard to do in my head with certainty. After I've slept on it, I might have to break out the pencil and paper.
::So, yeah, I've set a lower-limit to L, for various Ns, and can construct a ''possible'' upper-limit to L, but I haven't even checked these L(N)s vs. the values stated in the comic. Or what progress (and more advanced logical reasoning) has already been made in the field. I suspect I'm just reinventing the (hyper-)wheel, of course, rather than have the key to the problem that everyone else had failed to spot, but that's not the point. If I get even half way close to what the 'professionals' in this field have managed, I'll be smug and self-satisfied enough for myself. And, anyway, I've explained myself enough tolet any ''other'' similarly-minded nerd the ability to get at least as far as I've got with this problem. Which is as good an outcome, as far as I'm concerned, as getting this done entirely on my own. [[Special:Contributions/82.132.244.41|82.132.244.41]] 00:33, 7 August 2025 (UTC)
::: I'm still having trouble getting hold of long enough snakes. [[Special:Contributions/82.13.184.33|82.13.184.33]] 08:31, 7 August 2025 (UTC)
::: I've got the 50 length one for 7D, lets see if I can go further :) --[[User:Darth Vader|Darth Vader]] ([[User talk:Darth Vader|talk]]) 13:25, 7 August 2025 (UTC)

Psychology is way ahead of y'all, they've been putting actual mice in weird boxes for ''decades''. [[Special:Contributions/177.12.49.23|177.12.49.23]] 22:45, 6 August 2025 (UTC)
:Psychology might have been putting animals in boxes for decades, but zoology has been doing it for centuries! [[Special:Contributions/97.118.209.207|97.118.209.207]] 00:36, 7 August 2025 (UTC)
::Gastronomy has been doing it for as long as people have been storing food. [[User:BunsenH|BunsenH]] ([[User talk:BunsenH|talk]]) 03:41, 7 August 2025 (UTC)
:: https://scratch.mit.edu/projects/284912743/--[[Special:Contributions/2001:4450:8178:2200:D1C2:8DED:F6FE:E93C|2001:4450:8178:2200:D1C2:8DED:F6FE:E93C]] 04:01, 7 August 2025 (UTC)
:::That link doesn't work. When I remove the -- at the end it goes to some kind of math game. [[User:Barmar|Barmar]] ([[User talk:Barmar|talk]]) 15:58, 7 August 2025 (UTC)

Reading (just) the [comments] of the underlying research suggests that 98 is the longest found snake. Perhaps that means a longer one has not been explicitly eliminated (making 8 also not solved to some extent) [[Special:Contributions/2A02:A45B:8867:0:BED8:F2BA:838E:765|2A02:A45B:8867:0:BED8:F2BA:838E:765]] 22:52, 6 August 2025 (UTC)
:a(8)=98 was proven by: Östergård, P.R.J., Pettersson, V.H. Exhaustive Search for Snake-in-the-Box Codes. Graphs and Combinatorics 31, 1019–1028 (2015). https://doi.org/10.1007/s00373-014-1423-3
:[[User:Zmatt|Zmatt]] ([[User talk:Zmatt|talk]]) 09:46, 7 August 2025 (UTC)

I suppose Randall doesn't consider [[beetles]] cute, or else [https://en.wikipedia.org/wiki/Philosophical_Investigations#Wittgenstein's_beetle philosophy of language] would be included. [[Special:Contributions/137.25.230.78|137.25.230.78]] 23:15, 6 August 2025 (UTC)
: that's a great example [[Special:Contributions/177.12.49.23|177.12.49.23]] 01:46, 7 August 2025 (UTC)
: very nice one, reminds me of the story of the bug (moth) and debugging in computer science

In simultaneous interpreting, humans are the cute animal in the box. {{unsigned|DrInterpreter|07:35, 7 August 2025 (UTC)}}

I don't believe an explanation of the Schrödinger's cat thought experiment is necessary to understand this comic. However, people keep editing the page to include an incorrect description of the experiment, by saying the cat is either dead or alive and you don't know which until you open the box. That's wrong and misses the point of quantum superposition. The cat is not dead or alive, it's literally both, due to its fate being linked to radioactive decay, a process that is subject to quantum superposition. Since it does seem inevitable that someone will keep editing this to add an explanation, I've added one myself. [[Special:Contributions/177.12.49.23|177.12.49.23]] 10:29, 7 August 2025 (UTC)

The link in the mail newsletter lead to "http://https//xkcd.com/3125/", not sure if that's worth documenting here. [[User:Fabian42|Fabian42]] ([[User talk:Fabian42|talk]]) 13:07, 7 August 2025 (UTC)

Not a chemistry grad student, but is it possible that Randall intended "lure campus squirrels into laundry hampers in the hope that it ''sparks'' inspiration" as a humorous method of investigating the triboelectric effect? [[Special:Contributions/129.222.87.163|129.222.87.163]] 13:25, 7 August 2025 (UTC)

n=2: Snakes on a plane. [[Special:Contributions/64.201.132.210|64.201.132.210]] 16:47, 7 August 2025 (UTC)

the comic number 3125 is 5^5 [[Special:Contributions/96.77.127.105|96.77.127.105]] 18:26, 7 August 2025 (UTC)

Conclusion: we better have no snakes in the world {{unsigned ip|102.117.215.0|08:21, 8 August 2025 (UTC)}}

In the middle of the tree "wrong" examples, is the vertex leading from head to tail also a "wrong" vertex and should be coloured red? [[User:IIVQ|IIVQ]] ([[User talk:IIVQ|talk]]) 06:01, 9 August 2025 (UTC)

I suspect the squirrel fur + hamper + spark implies something electrochemical, though I haven't quite made the connection between the animal and the box.