Pangram | Hacker News Reader

Pangram

en.wikipedia.org

148 pointsbryanrasmussen64 comments

For anyone that plays Wordle on NYT Games site, there is another anagramming game called Spelling Bee in which the notion of a pangram is featured prominently. Basically, you get a set of letters from which to build anagrams (one privileged letter must appear in all your anagrams). Pangrams are considered to be any anagram which makes use of all the given letters and it will flash the word 'Pangram' on the screen when you get them, to give you a minor nerdy dopamine hit.

I am addicted to the game and get pissed at myself if I don't get to the top scoring tier every day.

EDIT: Removed spoilers for future readers.

Love that game, addicted too! ps. there’s a fourth pangram today you missed :)

Thanks! I love it so much. The one thing that annoys me is that it rejects anything that it considers to be domain knowledge or some such criteria. For example, today you'd think you could use 'potto', which is a cool animal in the loris family. But nope, Spelling Bee seemingly rejects it because it considers it to be specialized zoological knowledge or something. BS.

EDIT: Removed spoiler.

Psst, there are actually five.

!! thanks now furiously searching :D

Think "employment contract."

Yeah, +1 to Spelling Bee -- that game is great!

That said, this post needs a SPOILER tag.

has anyone here managed to get all the words after the genius level? What happens then? I just missed 3 yesterday

When you get them all, it's "Queen Bee." I've only ever gotten it once -- for whatever reason, the puzzle just clicked that day -- but it was extremely satisfying.

Good lord, I play every day and didn't know this was even a thing. I usually quit as soon as I hit Genius. I clearly need to level up!

Praising the game this much and not posting a link must be some sort of cruel punishment.

https://www.nytimes.com/puzzles/spelling-bee

I play that every day. I do my best and once I've given up by the evening time (it resets at midnight Pacific), I go to this site and see how much better I can do with the prefixes. You can tweak the checkboxes to show you more or less, all the way up to revealing the answers. Great site.

https://www.sbsolver.com/nctx/latest

If you happen to speak Icelandic (whopping 0.004% chance!), I recently made a clone of this game to learn NextJS: https://skrafl.rkl.is/

2 biggest differences being:

- Unlimited playing requires no registration.

- The answer keys are not just chilling there in the source code.

Inverse pangram challenge? This string is a frequency map of English letter usage in words (relative to the least common, q = 1). How many different sentences can be constructed from it, using all letters once (using say, the scrabble dictionary as the list of acceptable words)?

'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaabbbbbbbbbbbccccccccccccccccccccccccddddddddddddddddddeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeffffffffffggggggggggggghhhhhhhhhhhhhhhhiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiijkkkkkkllllllllllllllllllllllllllllmmmmmmmmmmmmmmmmnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnooooooooooooooooooooooooooooooooooooopppppppppppppppppqrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrssssssssssssssssssssssssssssssttttttttttttttttttttttttttttttttttttuuuuuuuuuuuuuuuuuuuvvvvvvwwwwwwwxxyyyyyyyyyyzz'

I have a feeling this blows up pretty fast, computationally.

> I have a feeling this blows up pretty fast, computationally.

Of course it does, but there's no way to compute the number without checking all combinations. There are hundreds of thousands of English words, and about the only initial filtering you can do is remove words with more than one q. Each partition will contain at least 20 words, so have >20! permutations.

Define sentence. Does it need to have meaning? Do sentences have to follow each other (or can they be unrelated)?

I wrote a multi-threaded anagram engine some years ago in Rust. It uses dynamic programming to linearize the problem as much as possible. The word list I typically use it with is not comprehensive by any means. I've taken words out, for instance, that would upset people to see in anagrams of their names (not fart, say, but you can imagine the sort of thing). Even with a simple set of word counts the number of anagrams, excluding permutations, of even simple things is vast. I get 6 from my handle here, which itself isn't very promising. A challenging short name like "Elon Musk" (it has a "k"), gives 36 (examples, "sulk omen" and "elk muons"). Things blow up quickly. My own name, including the middle name, has 23 characters, none of them difficult. The program finds 37,277,770, again, not including permutations. So with your character counts I think you're looking at a pretty big number. And you can blow it up further with a bigger, less polite word list.

>I've taken words out, for instance, that would upset people to see in anagrams of their names

appreciate this. sincerely,

Lana Ginger Shoales-Watts

In Sanskrit, where poets have endlessly delighted in playing such tricks with language, there are at least a couple of such verses that contain all the consonants in order: one composed in a live performance by a poet who lived 1854–1914, one from the 11th century or earlier (https://linguistics.stackexchange.com/a/21366).

It's a fun programming/puzzle challenge. I developed a perfect pangram in Dutch about a teachers bike that is very lightweight and fast, but not so strong

Jufs BMX: hypervlot c.q. zwak ding

"C.Q." is an abbreviation that could mean different things depending on the context. It is not a widely recognized or commonly used abbreviation, and its meaning is not immediately clear. Can you provide some more information or context about it?

It is common in Dutch, short for casu quo (latin).

"A c.q. B" means something like "A or otherwise B" (example translated from Dutch wikipedia page)

https://nl.wikipedia.org/wiki/Casu_quo

sorry didn't see anything in english on that page.

it's strange how a language that absorbs everything it comes across didn't bag this, perhaps there something else in english that does the same job?

In amateur radio, especially CW/morse code, CQ means "looking for someone to talk to". It's an abbreviated homonym for "seek you".

If you hear "CQ CQ CQ de WB6NOA" being transmitted, it would mean the person who owns call sign WB6NOA is looking to talk to someone on this frequency.

i did know that (in the back of my brain, and watching Ellie in "Contact") but don't see how it applies here.

Apologies, I meant to respond to another message.

In Dutch, c.q. is an abbreviation of the Latin term casu quo, meaning "or instead, alternatively" (lit. in which case). It doesn't really fit here because it suggests that vlot (quick) and zwak (weak) are interchangeable adjectives -- nor does it match the English usage of the same term, where it is used more like a premise/supposition rather than a conjunction.

Another populair one is: "Lynx c.q. vos prikt bh: dag zwemjuf!"

Nymphs vex, beg quick fjord waltz.

(27 letters, spotted in the classifieds of a free rag in DC back in the 70s or 80s)

Out of curiosity, I asked ChatGPT to generate 10 new pangrams.

The list it provided certainly had the feel of some of the more vexing examples of a pangram, but curiously did not contain all 26 letters.

I then asked it to count how many unique letters of the alphabet each contained. It (incorrectly) asserted that each line had all 26 characters.

I’m not surprised, just disappointed.

I asked it to produce a perfect pangram, which uses each letter once and only once. It unsurprisingly couldn't do it despite explicit instructions. It did produce several known examples of pangrams using all 26 letters but didn't seem to have "Mr Jock, TV quiz PhD, bags few lynx" memorized.

Thought this was really interesting, the "Thousand character classic"

>In Chinese, the Thousand Character Classic is a 1000-character poem in which each character is used exactly once...

Suppose that someone submitted a pull request that remedied a problem in the Coq proof assistant in a somewhat-too-clever way, but it was nonetheless approved by Coq developer Jim Fehrle.

In that case, we might say

  Zany Coq bugfix reviewed (thank Jim, please!)

Marcin Ciura has famously discovered many new Polish pangrams:

https://marcinciura.wordpress.com/2018/06/24/perfect-polish-... https://marcinciura.wordpress.com/2018/12/17/more-perfect-po...

And submitted an IOCCC entry in 2019 on the topic:

https://www.ioccc.org/2019/ciura/hint.html

some pangram heterograms found by others:

  Blowzy night-frumps vex'd Jack Q.
  Cwm fjord-bank glyphs vext quiz.
  Squdgy fez, blank jimp crwth vox!
  Q-kelt vug dwarf combs jynx phiz.
  TV quiz drag nymphs blew JFK cox.

program i wrote that found 184000 pangram heterograms in the scrabble dictionary:

  https://github.com/jkuli-net/pangram

> Other writing systems may present more options: The Iroha is a well-known perfect pangram of the Japanese syllabary

You still occasionally come across dictionaries ordered by this poem.

In word (and probably all word processors with proper Japanese support) you can make bullet list that follows this poem instead of going 1-2-3 or A-B-C.

Back in my Apple fanboy days, I was delighted to come across a pangram referring to Jobs and Wozniak, but I can no longer quite recall it, and my Google Fu is failing me. I think it was:

Jobs and Wozniak quibble mightily over expensive coffee.

It checks out, but it feels weird that I can't find a reference to it online.

When the Wordle craze happened, I wanted to come up with the best four word combo that hit as many common letters as possible. Not quite a pangram, but pretty close, and as memorable as possible.

"DWARF CLUBS MOPEY THING". All it's missing are JZXQVK.

Nice! I've been using AUDIO as my opening word which will give me four out of the five vowels.

I find vowel heavy is sometimes not great. Hitting a semi obscure letter like b or c is much more helpful than e.

For a three word combo, CLAIM, DROWN and SHUTE cover the 15 most common letters with no overlaps.

If you're interested in pangrams, you might like Dave's Scrabblegrams: https://twitter.com/dc_scrabblegram

Use each tile from a standard Scrabble set once to make a meaningful sentence.

For example:

> Oi oi, fallible developer! I, idiot programming enthusiast, frequent Hacker News, mixed subject extravaganza, do you? No way!

This uses each Scrabble tile once (with blanks M and X).

back in the day I ran a fun ruby quiz to make pangrams using the posix utilities: http://rubyquiz.com/quiz97.html

How long before someone makes a pangram wordle-like game?

https://www.scientificamerican.com/article/computer-recreati...

it's dewdney, but paywalled

Ah the page is back!

See: https://en.wikipedia.org/wiki/Wikipedia:Articles_for_deletio...

It gets deleted periodically for Wikipedian reasons like "It is mostly comprised of nonsense phrases thought up by people who apparently find this sort of thing terribly clever."

Interesting haha, looks like it's translated to 55 languages though!

The deletion of wikipedia articles makes no sense to me and always feels to me like a random editor going on a power trip

I routinely try to look something up that I remember had a page 5 years ago but now redirects to some 'list of x' page where instead of having a full article about it, it is just a bullet point with no description. Why delete the article? Definitely not to make the site better...

that page "List of pangrams" was deleted. The OP is a link to the Pangrams article which contains the contents of the deleted page.

I asked OpenAI for the shortest pangram in the English language. This is its response:

> The shortest pangram in the English language is "The quick brown fox jumps over the lazy dog."

Clearly it didn't read the wikipedia article.

Then I asked for an "original pangram":

> "Fjord zoologists quip jovially, waxing lyrical about xanthic lutrines."

A Google search indicates that it may, in fact, be original.

>> "Fjord zoologists quip jovially, waxing lyrical about xanthic lutrines."

> A Google search indicates that it may, in fact, be original.

Also distinctly non-pangrammatic

Maybe, but it's not a pangram. ChatGPT is slippery. (There's no "k". I stopped looking there.)

There's a "b" in "about", but there's no "k" or "m".

No 'm' either:

    (def phrase "Fjord zoologists quip jovially, waxing lyrical about xanthic lutrines")
    (require '[clojure.string :as str])
    (-> phrase str/lower-case distinct sort)
    ;; => (\space \, \a \b \c \d \e \f \g \h \i \j \l \n \o \p \q \r \s \t \u \v \w \x \y \z)

(edit: OP originally just mentioned 'k' but then ninja-edited in the 'm')

You are right, I did - sorry. I wrote something similar in Ruby to check!

Python:

  >>> phrase = "Fjord zoologists quip jovially, waxing lyrical about xanthic lutrines"
  >>> import string
  >>> set(string.ascii_lowercase) - set(phrase.lower())
  {'k', 'm'}

Concise!

It was the best of times, it was the BLURST of times! You stupid monkey!"

It's not slippery, it's just that it is not constrained to only make claims it can verify (and it's not obvious it's even aiming at validity/truthfulness in the first place).

And those are the sort of words that are rich in unusual letters and therefore likely to appear in pangrams in the training data.

I think I know where the bug is: there needs to be 26 unique characters. The first letter is capitalized and there's a space.... so technically, 26 ASCII chars - but not all 26 letters of the alphabet.

ChatGPT seems to give incomplete pangrams. It's also bad at counting syllables in poetry. I think it is just bad at counting; much room to improve here.

Reply on news.ycombinator.com