The Nineteenth Byte

Luis Mendo

12:00 AM

Did your dad already confess it was him who broke the server? :-P

Pavel

My dad works at Microsoft. There are thousands of potential suspects.

ATaco

As an Adult, this is exactly what adulting is.

El'endia Starman

Not long after my company moved into a new office, our internet went down for at least half an hour. It's amazing how little a company full of software engineers can do without internet. We were sent home, and since then, improvements were made to make sure we would never have our internet go out (unless AWS goes down or something, which has briefly happened).

HyperNeutrino

12:21 AM

Do you think that posting an index/challenge link list of all of the advent challenges to meta PPCG would be on-topic for PPCG? I think so since it's about PPCG itself, but I'd like to ask here before posting (asking on meta would probably work better but it seems kind of strange to ask on meta whether or not I can post a thing to meta :P)

ATaco

12:47 AM

meta-meta applies to meta.

HyperNeutrino

ok :P but like still I don't think it's worth posting a meta post just to ask if I can post a singular meta post

seems kinda pointless, ya know :P

Pavel

1:24 AM

> On the first GET of /christmas, the server sent to me
A few lines of plain, static text.
---------

On the second GET of /christmas, the server sent to me
Two tracking cookies
And a few lines of plain, static text.
---------

On the third GET of /christmas, the server sent to me
Three style sheets
Two tracking cookies
For a few lines of plain, static text.
---------

On the fourth GET of /christmas, the server sent to me
Four megs of bundles
Three style sheets
Two tracking cookies
For a few lines of plain, static text.

Christopher 2EZ 4RTZ

I did super good on my calc for once :P

El'endia Starman

@HyperNeutrino Yeah, it's usually better to just ask the question instead of asking if you can ask a question. In this case, however, what you want to post is not a question. Hence, I'd say it's probably a good idea to ask Meta about the possibility of making an Advent Challenges meta post.

HyperNeutrino

Ah okay. This might actually be applicable to future uses as well in case someone else decides to make a challenge series :P

ATaco

Honestly, this server needs to more relyably send back data.

Adám

1:42 AM

@Pavel Source?

Pavel

@Adám reddit.com/r/ProgrammerHumor/comments/7hix0f/…

I fixed a couple bits

Steven H.

2:28 AM

I like the first GET a lot more than the last... I thought it's supposed to get better ;-;

New Meta Posts

0

Q: Challenge Links for Challenge Series: On-Topic for PPCG?

As some of you may already know, there is currently a challenge series called the Advent Challenge series, posted entirely by me (so far). The purpose of this meta question is to determine whether or not posting a link list would be on-topic for posting to meta. For example, I would post a brie...

discussion

New Main Posts

3:12 AM

0

Q: The Missing Digit

Question: Given a string of random integers from 0 to 250 exclusive [0, 250) with one digit missing, find the missing digit. That is, a string with 0 delimiters (something like "23,54..."; is instead "2354....), and 250 numbers, there is one random number missing. Test data is included. This is...

code-golf string number sequence parsing

Riker

3:27 AM

even github is joining the net neutrality protest

@Pavel lost it at "FIVE GOLDEN PINGS"

Riker

4:13 AM

@Pavel doorknob & alex have been inactive for a while now, they were pretty good mods

I don't think you were there for when they were active

Pavel

@Riker I was (for Doorknob, not Alex), but what I said is meant to reflect that.

Riker

hm, ok

Pavel

Mind you, if anyone should be a mod, I think it should be Mego.

Riker

or DJ

Downgoat

if elections were today I would probably vote to replace Alex&Doorknob with DJ&Mego

though I did already vote for mego last election

Pavel

4:18 AM

@Riker DJ too, absolutely. But if I had to pick one, it'd be Mego.

Downgoat

@Riker they 'joined' a long time ago but it's just the banner that's new

Riker

@Downgoat well yeah

@Pavel hm, ok

Anonymous

@Pavel I'm flattened. Wait, no the other one. Flattered.

orlp

@LeakyNun I'm off to bed right after I write this message

but here's a little challenge for you

a(1) = 1, a(n+1) = (a(n) + 3) / (a(n) + 4)

Downgoat

School: pffffttt you don't need to know about recycling. Leave everything in a hole.
School: Y U NOT KNOW WHICH ARTIST LOUIS XIV WAS PATRON OF

orlp

4:22 AM

find a closed form for a(n)

Anonymous

4:43 AM

@orlp I decided to give it a shot. I got KDErMyoqKG4tMSkpLygxKzQqKihuLTEpKQ== (base64-encoded for spoilers)

Thomas Ward

@Mego you shouldn't have said it was encoded.

should have made people figure it out themselves :P

Anonymous

I mean I encrypted it. You have to make a GUI in Visual Basic to decrypt it.

Anonymous

@ThomasWard The == at the end kinda gives it away :P

Thomas Ward

@Mego not really... not if you use a really obscure encoding :P

Anonymous

Seeing = or == at the end of a string of ASCII alphanumeric characters is a dead giveaway that it's base64-encoded. Even if it isn't, base64 is the first thing to try.

Pavel

4:48 AM

Are there any easy to use encryption schemes in coreutils that aren't as obvious as base64

Thomas Ward

@Pavel expand it to be non-coreutils but instead the Python3 base64 library and then that becomes a "yes"

Pavel

@ThomasWard Howabout "Something I probably already have installed and is easy to use from the command line"

Thomas Ward

do you have python3 in your system?

Pavel

Sure, but a python -c with multiple statements isn't easy to type

Thomas Ward

mm, true.

ascii85 is a fun encoding though :P

C^0KDDk>;$Eiox6FDNl9G%6}6ZY?n>DF should be equivalent to @Mego's b64 encoded string in decoded form

Pavel

4:54 AM

That is better. No ==.

Thomas Ward

yup

tio.run/##BcHLDYAgDADQVerJkqAXPyFxATfwbLVGEqQE6sHp8b306S1xqNU/… if you're bored and want to experiment.

Pavel

6:09 AM

I just bought something with bitcoin for the first time. I felt like a wizard for about 15 seconds before waiting for the transaction to go through for an hour.

ATaco

The "How to juic Avocad" question never actually had the text "Avocad" in it.

user202729

6:26 AM

Can some moderator clean up the comment section in this question‌?

user202729

6:47 AM

^ Well, it's deleted now.

Sherlock9

7:40 AM

Hello, I'm in a bit of a pickle

My undergraduate thesis proposal is due 23:59 of the 15th

And my case study doesn't have a case to study :')

Do you know of a system that is prone to relatively complex bugs

Or even a language parser or KotH controller that could use some checking?

My idea is to use TLA+, a formal specification language, to test if the system is correct (for my purposes, without bugs), but I'm having trouble finding that software system. Preferences are systems that deal with concurrency and stateful messes

Anonymous

7:54 AM

@Sherlock9 Pick any webserver :P

Anonymous

In fact - use PHP

Sherlock9

@Mego Got one in mind? :D

ye gods

That may actually work

And kill me

Anonymous

PHP is synonymous with mess

Sherlock9

The thing is, since this is based on its own language and based on the specification of the system, but not the code itself, PHP may not actually come into play

The bad habits it breeds like rats will

... Best of both worlds really :D

Still do you have something in mind that I can poke around in?

Anonymous

Umm

Sherlock9

7:58 AM

I'm going to have to be relatively specific in the proposal, unfortunately

Anonymous

php.net/manual/en/features.commandline.webserver.php

Anonymous

Or for a more classic example, you could look at pthreads in C

Anonymous

Or for a more modern example, the PPCG v2 codebase :P

Sherlock9

I was actually about to ask sheepish smile

Anonymous

For a while, not having imports in a specific order would cause all sorts of hell to break loose. I fixed that so that imports are mostly independent of the order, and only minor hells break loose when things go wrong.

Sherlock9

8:02 AM

Is it okay if I use the codebase for my thesis though? I think I only need to the code that I wrote for the testing into the proposal, but I'm not sure what license it's all under

Anonymous

Now the SQLAlchemy ORM is giving me a headache

Anonymous

@Sherlock9 Sure. It's licensed under MIT, so go nuts (as long as you give us credit).

Anonymous

Also if you find any errors, let us know :D

Sherlock9

Oh absolutely!

On both counts

On the credit, I'm required by law :D

On the errors, well that's the whole point

Anonymous

And if you have any questions about our mess, ask us in the chatroom

Sherlock9

8:03 AM

It's mathematical debugging :D

Thank you a million times :D

Anonymous

No worries - it's benefitting us, too, so everyone wins

Anonymous

in Charcoal HQ, 25 mins ago, by Processed Meat

-2

Q: WorldEyeCam is a CCTV Security products distribution company?

We proudly offer 1st class technical support for our customers we serve with the complete understanding that it is you, the customer that matters to every successful business. If you haven't already learned the hard way, you will understand that it takes more than just a website and the ability t...

discussion careers area51 stackexchange spam

Anonymous

A spammer posted spam on Meta.SE with the spam tag

3

Leaky Nun

9:25 AM

@orlp letting the n-th term be ((a_n)x+(b_n))/((c_n)x+(d_n)) gives us 4 linear mutually recursive sequences, which shouldn't be hard to solve

where x is a0 unevaluated

that even works for any starting value a(0)

Anonymous

@LeakyNun That is certainly much more difficult than other approaches

Leaky Nun

@Mego 1. but more general :P 2. just a proof that I can do it

Anonymous

Induction is quite general, too :P

Leaky Nun

but you need to have a closed form before you can do induction

Anonymous

Err, not induction - I don't know what the formal name for it is, but my prof always (incorrectly) called it an inductive approach

Leaky Nun

9:33 AM

then your approach is just mine with x=1

and you should have 2 mutually-recursive sequences instead of 4

Anonymous

I didn't use any linear algebra to solve it - just simply recognizing a pattern

Leaky Nun

p(0)=1, q(0)=1, p(n+1)=p(n)+3q(n), q(n+1)=p(n)+4q(n)

and then a(n) = p(n)/q(n)

Anonymous

That's not at all what I did

Leaky Nun

but it works :P

[p(n+1);q(n+1)] = [1,3;1,4] [p(n);q(n)]

Anonymous

Sure, but it's a lot of extra work

Leaky Nun

9:37 AM

(x-1)(x-4)-3 = x^2-5x+1 = 0 --> x=(5+-sqrt(21))/2 rip

(1,1),(4,5),(19,24),(91,115)... I don't see any pattern

Anonymous

Go back to the original definition. a(n+1) = (a(n) + 3)/(a(n) + 4)

Leaky Nun

1 - 1/(a(n)+4)?

Anonymous

Nope, slow down

Anonymous

Give me a recursive definition for a(n)

Leaky Nun

you've just quoted it

Anonymous

9:40 AM

That's the definition of a(n+1)

Anonymous

Err

Anonymous

I typed that wrong

Anonymous

We can say a(n) = (a(n-1) + 3)/(a(n-1) + 4)

Anonymous

Now define that in terms of a(n-2)

Leaky Nun

so a(n+2) = (4a(n)+15)/(5a(n)+19)

and we're back to the sequence

Anonymous

9:43 AM

No...

Leaky Nun

4 mins ago, by Leaky Nun

(1,1),(4,5),(19,24),(91,115)... I don't see any pattern

I mean, a(n) = (4a(n-2)+15)/(5a(n-2)+19) if you like this better

it doesn't make any difference

Anonymous

Now do a(n) in terms of a(n-3)

Leaky Nun

you'll just get the sequence back

Anonymous

Oh, I see what you mean by earlier

Anonymous

Ok

Anonymous

9:46 AM

Wait how on earth did I arrive at my solution when I did it completely wrong?

Leaky Nun

wait

a(n+1) = 1/(1+1/(a(n)+3))

we start with 1

a(1) = [0;1,4]

a(2) = 1/(1+1/(3+1/(1+1/4))) = [0;1,3,1,4]

wait we aren't finding limits though

the limit would be [0;(1,3)]

and it does have a limit

Anonymous

So my logic was flawed in how I came to my solution, but somehow the solution is still correct

Leaky Nun

the limit would just be the solution of x=(x+3)/(x+4), i.e. x^2+3x-3=0

@Mego ????

Anonymous

I did my substitution wrong, but still got a valid solution somehow

Leaky Nun

how do you know your solution is valid?

Anonymous

9:51 AM

Because I tested it

Leaky Nun

...

Anonymous

It lines up with the recursive solution for n in [1..1000]

Anonymous

Working on a proof by induction now

Anonymous

Oh I'm an idiot

Anonymous

My solution only seemed correct because I was accidentally using integer division

Anonymous

10:02 AM

And the sequence is always <= 1 so integer division was hiding the differences

feersum

10:46 AM

Non-integer sequences?

Blasphemy.

Luis Mendo

11:13 AM

math \o/

Sherlock9

@LeakyNun I do

Starting from the second term, the difference between the numerator and denominator of a_n is the denominator of the last term.

Sherlock9

11:36 AM

@LuisMendo Congratulations!

orlp

11:48 AM

@LeakyNun the limit is easier yes

orlp

12:02 PM

@Mego I don't believe that's correct.

@LeakyNun yes, now solve it :)

user202729

The formula given by Mathematica is very complex, I don't think I know of a way to solve it.

Luis Mendo

12:18 PM

@Sherlock9 Thanks!

orlp

@user202729 I know two ways :)

Luis Mendo

12:32 PM

The "inbox" and "achievements" links on the nav bar stopped working for me today. For example, I see a red circle "1" indicating I have a notification, and clicking on it does nothing. I'm using Chrome on Windows 10. Anyone else?

Mr. Xcoder

2 days left to save the OEIS challenge... Sorry for giving such a terribly complicated sequence

Luis Mendo

(It works on Edge)

Mr. Xcoder

@LuisMendo Works fine on Safari

Luis Mendo

Weird, It worked for me yesterday, same system

Well there was a Windows update...

Mr. Xcoder

And BTW Congrats for math!

Luis Mendo

12:34 PM

Thanks :-)

J. Sallé

@LuisMendo Mine are working properly on Chrome 62 + W10, but my windows didn't update yet.

Luis Mendo

@J.Sallé I'll restart during lunch time and see if it helps

New Main Posts

12:55 PM

0

Q: Solutions to Google's doodle on kids programming

Today's Google doodle is about Celebrating 50 years of Kids Coding: The goal is program the path of a little bunny so that it can eat all the carrots. For instance, this is level 4: Green pieces = go one step forward Blue pieces = turn left/right (but stay on the same block) Orange pieces = lo...

golfscript

user202729

1:18 PM

@Mr.Xcoder That's usual, and definitely not the first time. Me again, or @NieDzejkob or @HyperNeutrino or ChristianSiever ? Or somebody else?

@TFeld Does it? It's a question about (atomic) golfing in a very domain-specific (very much non-turing complete) programming language. — Martin Ender ♦ 13 mins ago

Is it really non-Turing-complete?

Martin Ender

If it has any chance at being TC, you need to consider the level layout part of the program.

Or maybe the input, I guess.

Basically, the only way you can express a branch is to have the bunny run into the edge of the level.

New Main Posts

0

Q: Tree Traversal Problem (maybe)

My apologies in advance if this problem has a well known solution and it belongs in stackexchange. Maybe the answer is obvious but if it is, it is not obvious to me. What I need to do in the above is I need to loop through the lists. Think of the numbers as hypotheses. If the list has only on...

tree-traversal

NieDzejkob

@user202729 It's simple once Hyper fixes his A000109 answer :P

user202729

Huh? Which answer?

HyperNeutrino

1:40 PM

@user202729 There's apparently an error in one of my OEIS challenge Proton answers about 60-70 answers back and I have to fix it or else the chain breaks 60-70 challenges ago :P I can probably fix it, but I think there's a backup Mathematica answer posted by KSmarts in case mine broke. I'm not sure whether or not it was that one.

salty kid is trying to roast us all with their excellent lack of logic

2

One OEIS after another

Transcript for