Random thought of the day
-
What if the "Mana from heaven" was actually the fur sheddings of an alien that looked a lot like a cloud?
-
Don't make me reboot you again!
-
I think this architect liked Legos.
-
-
@TimeBandit
E_NOT_LEGO_BRICK
(but they do look pretty cool)
-
@dcon said in Random thought of the day:
@TimeBandit
E_NOT_LEGO_BRICK
(but they do look pretty cool)I keep thinking of that Heinlein short story.....
-
-
@da-Doctah said in Random thought of the day:
In the '80s I remember stumbling across people on eBay
That would be very impressive, given that eBay wasn't founded until '95.
-
@abarker said in Random thought of the day:
@da-Doctah said in Random thought of the day:
In the '80s I remember stumbling across people on eBay
That would be very impressive, given that eBay wasn't founded until '95.
It was the late eighties.
-
@Tsaukpaetra said in Random thought of the day:
What if the "Mana from heaven" was actually the fur sheddings of an alien that looked a lot like a cloud?
TIL @Tsaukpaetra does drawing.
-
License plate thought of the day: FDN: a new dynamic network management protocol invented by @Polygeekery
-
How does this protocol interact with firewalls?
-
@Zerosquare said in Random thought of the day:
How does this protocol interact with firewalls?
This protocol will be so hot that it will obsolete them
-
-
@dcon That must be a nightmare to sweep in wartime.
-
@acrow I doubt that's a significant consideration for most architects.
-
@Watson what is the thought?
-
Pick any number. It can be a postiive integer, a decimal fraction, or an irrational constant. Any number.
Write out the number using words.
Count the number of letters in the written name of the number.
Take that number of letters, and write it out in words, count the letters, and repeat.
Eventually, no matter what number you started with, you will hit the number 4, which has four letters, which means that subsequent steps will repeat "4, 4, 4..." indefinitely.
This works in English, and apparently in German, but not in Spanish (where you'll either hit 5 and repeat, or cycle endlessly between 4 and 6).
-
@da-Doctah At a glance, it looks like Greek should either repeat 5 (πέντε) or alternate between 4 (τέσσερα) and 7 (επτά).
-
Tried it in Polish for pięćdziesiąt osiem, siedemset dziewięćdziesiąt, cztery miliony osiemset dwadzieścia pięć tysięcy czterysta cztery, trzysta pięć, and czternaście.
The cycle is always 6->5->4->6. Neat.
Works for complex numbers too!
-
@Gąska said in Random thought of the day:
Tried it in
Polishkeyboard mashing with random squiggles for pięćdziesiąt osiem, siedemset dziewięćdziesiąt, cztery miliony osiemset dwadzieścia pięć tysięcy czterysta cztery, trzysta pięć, and czternaście.FTFP
-
The cycle for French is 6 (six) → 3 (trois) → 5 (cinq) → 4 (quatre) → 6 etc.
-
@HardwareGeek technically it works for any string of letters whatsoever (including empty string), in every human language. And although it's pretty easy to prove it by hand waving, I'd love to see a formal proof. I know it's possible, and I roughly know how to achieve it, but I'm too
to actually do it.
-
The WTDWTF theorem: no matter what you try to achieve, you'll always hit the
point eventually.
-
@Zerosquare said in Random thought of the day:
The WTDWTF theorem: no matter what you try to achieve, you'll always hit the
point eventually.
We'd prove it, but
-
Also know as "the proof is left as an exercise to the student".
-
@Zerosquare fuck you, I'm definitely doing it now.
What I figured out so far:
- After first iteration, you always end up with a natural number.
- The first few numbers can be proven on case-by-case basis.
- Due to composite nature of writing large numbers (e.g. "twenty ...", "hundred ..."), you eventually reach a number X such that every number Y larger than X can be written down in fewer than Y letters. Meaning that after enough iterations, you always end up at or under X.
- Once X is known, all that remains is proving the hypothesis for every number up to X one by one.
- You don't need to know exact value of X; all you need is upper bound.
- For English language, X is definitely 20 or less.
The only thing missing now is the formal proof of #3. I'll do it later.
...Damn you @Zerosquare!
-
(looks like your proof idea is the same as mine.)
-
By the way, I wonder if that would make a good interview problem? It'd make more sense than the "how many piano tuners could you fit in a Boeing 747" rubbish.
-
@Zerosquare I, for one, have used similar reasoning several times when solving hard problems at work. Semi-formal proofs that something will definitely happen a particular way if certain conditions are met are very useful when designing various algorithms. The downside is that you filter out everyone that isn't good at formal reasoning, which drastically reduces your candidate pool.
-
@Gąska
As far as #3 goes, natural numbers are effectively named by listing their digits (are there any natural languages out there that don't use a positional system?) modulo constant-sized grammatical flourishes; the list of digits grows logarithmically with the size of the number [makes handwaving gesture intended to demonstrate the difference between linear and logarithmic growth].
-
@Watson the hard part is saying that in the language of formal logic.
-
@da-Doctah said in Random thought of the day:
and apparently in German
Impossible, every German word is at least 12 letters
-
@Zerosquare said in Random thought of the day:
how many piano tuners could you fit in a Boeing 747
With or without their pianos?
Ah, nevermind
-
@Gąska said in Random thought of the day:
The downside is that you filter out everyone that isn't good at formal reasoning, which drastically reduces your candidate pool.
You consider that a downside?
Isn't filtering out unsuitables the sole purpose of interviewing?
-
@acrow AFAIK the point of interviewing is getting someone hired.
-
@Gąska said in Random thought of the day:
@acrow AFAIK the point of interviewing is getting someone who can add value hired.
FTFY
I'm not a business major, but I'm pretty sure that part's important for a healthy business.
-
@acrow said in Random thought of the day:
that part's important for a healthy business
You're excluding a great many places like that.
Also, it's mostly not that important for a particular person to be a great hire, at least at a larger employer. One of the advantages of size is that a bad hire won't usually sink the business, making hiring a much less risky activity overall. (Technically, all it takes for a business to be healthy is for income to be at least keeping up with outgoings, averaged over a few years depending on the creditworthiness of the business.)
-
@Gąska said in Random thought of the day:
@Watson the hard part is saying that in the language of formal logic.
Assuming you actually have a base-ten or whatever representation (it's not true for unary, but nobody uses that), all you need to do is basically show that the length of the base numbers (hundred, million, etc) grows slower than their size. And assume constant sized grammatical "glue". That's trivially done by exhaustion up to the point where you don't even know how to name new things anymore (what's 10^123 in non-scientific notation?).
-
@dkf I do understand that bigger businesses don't sink easily from one or two little mistakes or bad hires. But the point was that every hire is either an expense or a profit-maker (however hard it is to numerically determine the QA department's contribution to the company's profitability - and they do contribute, usually). And the interview's purpose is to determine which one a candidate is most likely to be. Profitable or costly? Every mishire will weigh on the bottom line.
For purposes of software engineering, capacity for "formal reasoning" (though not restricted to any particular notation - just the skill itself suffices) seems to me to be a rather foundational block. That is, based on my experience, persons incapable of that do not write productive software. So, I would not hire such persons for software roles. For other roles, maybe. But the above posts were discussing a formal reasoning -solved problem as a replacement for an "Interview 2.0" type of question - a Microsoft-introduced practise that mostly persisted in the software world only, AFAIK.
-
@topspin said in Random thought of the day:
Assuming you actually have a base-ten or whatever representation (it's not true for unary, but nobody uses that), all you need to do is basically show that the length of the base numbers (hundred, million, etc) grows slower than their size. And assume constant sized grammatical "glue". That's trivially done by exhaustion up to the point where you don't even know how to name new things anymore (what's 10^123 in non-scientific notation?).
I believe that'd be a quadragintillion.
-
@Zerosquare said in Random thought of the day:
The cycle for French is 6 (six) → 3 (trois) → 5 (cinq) → 4 (quatre) → 6 etc.
Hmm. In Portuguese most single digits' names redirect to four, except one tries to postpone by sending to two first, and four itself tries to redirect to six. So the most likely cycle is 4->6->4, I think.
Five is doing its own thing.
1 (um) -> 2
2 (dois) -> 4
3 (três) -> 4
4 (quatro) -> 6
5 (cinco) -> 5
6 (seis) -> 4
7 (sete) -> 4
8 (oito) -> 4
9 (nove) -> 4
-
-
I realized I couldn't remember the names of numbers in German after 5.
How could I have forgotten about sechs?
-
Māori has 11 (tekau ma tahi) -> 11; now I'm wondering if there is a language (not necessarily based in decimal) where a single-digit number has a name long enough to yield a two-digit number.
-
@Watson Yep, I think I nailed the "random thought" part of this thread pretty solidly this time.
-
https://cdn.discordapp.com/attachments/766584957984047134/769126232804818944/unknown.png
Status: Thinking about fucking mosquitos....
-
Pinging @error
-
-
Abstraction is pretty fundamental to almost all areas of software development.
With Node, they’ve finally managed to abstract from reality itself!