My first solution was RNG(hash(seed.toString() + string)), which would get rid of the correlation while still being deterministic based on the seed.
It's also more robust than calling RNG 10 times since if you use the same algorithm to seed as for the RNG proper then you will get the same sequences in each instance, just offset.
Fun game overall, took a little over 2 hours to beat. Some notes:
- I found the decade theming of the packs a nice touch.
- I'm mildly curious as to the maximum difficulty of the math problems.
- I did find the combat a little frustrating as the levels increased, I found that I had to basically ignore the block button if it was on cooldown at level 10 because the time between attacks is shorter than the cooldown, and I was definitely reaching the limit of my reaction times there.
- I recognised most of the movie quotes but some did elude me.
On the topic of incremental game gacha parodies, I couldn't help but think of the recent game One Trillion Free Draws (https://duducat.moe/gacha/), the premise of which is that you need to use your draws to get upgrades to use up your 1 trillion free draws faster. I've come across a few others in my time but I don't quite recall the others.
The 500 edits required for access to TWL is actually for all of the sites under the Wikimedia banner. If you're having trouble finding things to edit on Wikipedia, you can try their other sites such as Wikisource or Wikibooks.
On Wikisource in particular, it's fairly easy to make useful edits through validating proofread pages or proofreading simple pages (both of which can be easily found in the Monthly Challenge).
The conclusion is basically (Larry) Niven's Law of Time Travel. From "The Theory and Practice of Time Travel" (1971):
> If the universe of discourse permits the possibility of time travel and of changing the past, then no time machine will be invented in that universe.
The entire essay is worth a read, of course. Meanwhile the paper in the OP goes for a more mathematical approach.
Somewhat related is the work done in "Falling with Style: Factoring up to 255 “with” a Quantum Computer" published in the proceedings of SIGBOVIK 2025 [1]. The author, Craig Gidney [2], successfully factored all odd composite numbers up to 255 using Shor's algorithm, even though the quantum circuits involved were such that any meaningful output would be overwhelmed by noise (and indeed, performance was maintained when the circuits were replaced by a random number generator).
> To my knowledge, no one has cheated at factoring in this way before. Given the shenanigans pulled by past factoring experiments, that’s remarkable.
[2] Who has previous experience in cheating at quantum factoring: see "Factoring the largest number ever with a quantum computer", posted April Fools' Day 2020 at https://algassert.com/post/2000
A while back I came up with the idea to carve out 4096 code points in plane 14 (Supplementary Special-purpose Plane) for super-surrogates, and use three such surrogates (1 initial, 2 extension) for codepoints beginning from U+110000. If done properly you get unlimited range and self-synchronizing, at the expense of needing 12 bytes minimum per codepoint (more if you want it truly unlimited), but I figured the demand for UTF-16 would be low enough by the time it's needed that it's a workable tradeoff.
Having come across the connection from somewhere else entirely, I found it a very curious rabbit hole. I saw recently a website that attempts to summarise the whole thing (not quite complete) [1], with a section on St. Bride's.
Then there's the oral history someone posted on Something Awful some years ago [2] -- from someone who may have accidentally indirectly derailed the entire movement.
RFC 2550 to the rescue: handling the transition from A99999 to B100000, or even from Z999999999999999999999999999999 to ^A1000000000000000000000000000000 will be a piece of cake.
Neat. However, I noticed that the page description reads "No stress, memory-optimized 3D Tetris variant.", but I can't figure out what's supposed to be 3D about it.
MAX_SAFE_INTEGER = 2^53 - 1 = 9007199254740991. The next integer 2^53 is representable, while the following one 2^53 + 1 isn't.
2^53 + 1 milliseconds = 9007199254740993 milliseconds = 104249991 days 8 hours 59 minutes 0.993 seconds. Since the zero point is midnight 1970-01-01, we reach the first non-representable millisecond at 287396-10-12T8:59:00.993 (be careful to use the Gregorian leap rule if you want to calculate this manually).
In a similar vein, there is Wikisource.[0] Wikisource has the advantage of allowing for extensive formatting to closely match the source works due to its wiki-based format, but doesn't have quite as robust processes. Its flexibility is unparalleled though -- it covers virtually any form of scanned print work and even some old movies, and contributors can focus on whatever niches they're interested in if they want.
If you check the section "Extensions of the standard dictionary numbers" on the page you linked, there's a system for naming arbitrarily large numbers as -illions. The prime here is approx. 8.82 * 10^41024319, and 10^41024319 is the 13674772nd -illion (short scale), so it would begin eight tredecilliquattuorseptuagintasescentilliduoseptuagintaseptingentillion.
> I also read that starting with Aldous Broder and then switching to Wilson's Algorithm (reasoning: Aldous Broder is slow at the end, Wilson's Algorithm is slow at the start) is faster than either. However, I haven't seen proof that this combination still results in a uniform spanning tree (where all possible mazes have equal probability).
I did some searching, and the paper at [1] (2022) studies the problem. Based on the paper, a naive combination of the two algorithms can generate uniform spanning trees on complete graphs, but more work is needed for arbitrary graphs. [2] apparently cites this when discussing their hybrid maze generating algorithm, but I haven't been able to find a copy to check.
Funnily enough, there was a short story about the killion published in the New Yorker in 1982: "The killion, as every mathematician knows, is a number so big it can kill you." [1]
This reminds me of an old Roald Dahl story, "The Great Automatic Grammatizator". A man creates an electric computer that can write short stories in a few minutes (this being written in 1953, when electric computers were just becoming a thing), and starts churning out magazine submissions. Having made a name for himself as a writer (several names, in fact), he then upgrades it to write novels. Finally, he starts offering existing writers the chance to license their names to the machine's works. It's a good read.
Of course, the situation isn't quite the same, largely because unlike the story the magazine here can identify AI submissions, but the similarities are nevertheless interesting.
I once read a story in which a human gave an extraterrestrial visitor a calendar to examine and they were confused as to all the inconsistency. I then wondered how they would react if we gave them a Malaysian wall calendar instead:
- Four calendars (Gregorian, Islamic, Chinese, Hindu)
- Three scripts (Latin, Chinese, Tamil)
- Two solar (Gregorian, Hindu), neither synced with the other, with varying month lengths (29-32 for the Hindu, but at least months follow solar longitudes properly; Gregorian doesn't really follow anything)
- Two lunar (Chinese, Islamic), also offsync (Chinese starts months on new moon, Islamic on first visible crescent) and Chinese has leap months while Islamic doesn't.
- Weekday names in four languages (English, Malay, Chinese, Tamil)
It'd certainly be a puzzle for our hypothetical alien.
It's also more robust than calling RNG 10 times since if you use the same algorithm to seed as for the RNG proper then you will get the same sequences in each instance, just offset.