The Nineteenth Byte

quartata

12:00 AM

If you do bundling a header is way easier than bundling a shared object

George V. Williams

@orlp, you're right that's equivalent.

huh

except you need some details to distinguish when one integer stops and another ends

orlp

@GeorgeV.Williams not really

you can use variable base encoding in this case

or alternatively

arithmetic coding

with a uniform model of 0-2^32 for the first integer, then 0-i, which automatically reaches the shannon bound for it

quartata

@orlp I love how everything always seems to come full-circle to arithmetic coding for you :P

orlp

@quartata it's just so powerful!

having the model == having the encoding

George V. Williams

@orlp maybe I'm being dense, but how do you know what the variable base is?

like the first one is all 32-bits, but after that?

orlp

12:03 AM

@GeorgeV.Williams well, you know that for the first 'digit' the base is 2^32

then you extract that digit (d = n%2**32, n //= 2**32)

now for the next digit the base is d (I used i earlier, sorry for switching variables)

eh small correction

the base is d+1

considering that we might have twice the same number in a row

Pavel

@Downgoat can I delete your submissions to the time capsule?

ATaco

Do you have that spreadsheet of the submissions still?

Eᴀsᴛᴇʀʟʏ Iʀᴋ

@orlp you're like xnor but purple

you even more have 4 letter names with 1 vowel, o.

ATaco

His name is an Anagram for rolp. Rolp to my knowledge does not mean anything.

orlp

@ATaco no it's not

ATaco

12:15 AM

Not intentionally.

Destructible Watermelon

hi

orlp

@GeorgeV.Williams example implementation in Python

link to gist

>>> encode([5, 10, 13, 24], 50)
102687
>>> decode(_, 4, 50)
[24, 13, 10, 5]

ATaco

Needing to store 4, 50 kind of defeats the purpose.

orlp

no it doesn't

ATaco

Unless you're storing many of these lists.

orlp

12:19 AM

@GeorgeV.Williams the problem with this implementation is that it's not uniform output size

George V. Williams

@orlp, right I see.

orlp

>>> encode([50, 50, 50, 50], 50)
6765200
>>> encode([0, 0, 0, 1], 50)
1

George V. Williams

my idea was actually slightly different but I don't think it's any better

orlp

on average I believe it's optimal

but its worst case isn't

Luis Mendo

@Dennis Which spelling is (more) correct: Leibnitz or Leibniz? (the mathematician) My high-school teacher spelled it with the t, but I keep seeing it without

orlp

12:21 AM

@LuisMendo the latter

it's pronounced as the former though

El'endia Starman

@LuisMendo I've never seen it as the former. (But then again, I'm from the USA.)

Luis Mendo

@orlp So is it strictly incorrect with the t?

orlp

@LuisMendo after some googling it's not

it's french spelling

Luis Mendo

@El'endiaStarman Thanks. The option without t is gaining momentum :-)

orlp

z in german is pronounced ts

Luis Mendo

12:22 AM

@orlp So that may explain the change in spelling

Let's wait for Dennis though. He speaks German natively

orlp

but considering Leibniz was german

I think you should write Leibniz

Dennis

2

Luis Mendo

I think that's the safest. But I have fond memories of my professor and I'd like to spell it his way :-)

Dennis

That's his signature. No T.

Luis Mendo

@Dennis Hm that has not t, does it?

orlp

12:24 AM

@LuisMendo Lewis Mendo

that's what Leibnitz is :)

Luis Mendo

@Dennis Clarified then from a reputable source. Thanks!

Nice signature...

orlp

my signature is the worst

inconsistent

ugly

oh well

Luis Mendo

@orlp That's a nice analogy, yes :-)

@orlp Mine is simple. My name hand-written, basically

Dennis

Both spelling of the word Leibni(t)z exist in German. The mathematician and the cookie are spelt Leibniz, the city and the moon crater Leibnitz.

ATaco

What year was Leibniz?

Luis Mendo

12:27 AM

@Dennis So are they two spellings for the same name, or two different names? (If the distinction makes sense)

orlp

@ATaco Leibniz is a person, not an event...

ATaco

I am aware.

I need a time period.

Destructible Watermelon

I spawned in nethack and immediately I have magic marker

orlp

@ATaco google?

Dennis

@LuisMendo I'd say it's the same name.

ATaco

12:28 AM

As there is a time period in which many spellings of names, such as Shakespeare, are inconsistent even amongst themselves.

Downgoat

ATaco

I thought we finished bagging on Windows for now.

Pavel

@Downgoat can I kill your time capsule submissions?

Downgoat

D: oh no, why?

Pavel

You said that you weren't planning on implementing either of them

Downgoat

12:38 AM

I was joking

Pavel

When you said that or when you made the submission?

ATaco

Do you have the spreadsheet of submissions still?

Pavel

docs.google.com/spreadsheets/d/…

ATaco

Threead is going to be needlessly complicated and I love it.

Pavel

I'm just curious what Martin Ender is doing.

ATaco

12:54 AM

To add 1 to buffer a one must do:

    l
i1d
 i+d

Pavel

Can you explain that out?

xnor

i'm getting a blank winners by language column in the leaderboard of the RPS question -- could someone please verify?

Pavel

link the question?

No. loading fine for me.

ATaco

It's actually kind of simple, just know that each of those three lines, is a code running independingly, that can only write to its own memory tape.

the ` i1d` inserts 1 into it's memory tape, then removes it.

The ` i+d` Adds the 1 and whatever value was on the first tape, and deletes it immediately after.

But ` l` copies a value from the tape to the left of it, which is the value that was added just before it's deleted.

THis, thus, adds 1 to the tape.

The i and d are just there if you care about what's on tapes 2 and 3

Pavel

i1

?

ATaco

12:59 AM

i inserts a new cell into the memory tape at that position.

Which pushes other memory to either side.

d then deletes that cell. Allowing us to use the 1 and + only temporarily.

Pavel

I don't actually see any | in your code.

ATaco

Alternatively, the following would work.

(That's a lowercase L)

  l
1
 +

Pavel

l and 1 look similar

ATaco

Yes, they do.

I could change them to L and R

They simply grab the value on the tape to the Left or Right respectively.

Pavel

Nah, just because it's not great in this chat isn't a reason to change it.

ATaco

1:02 AM

Doesn't look good in Sublime Text either.

Pavel

Well, do what you want.

ATaco

The 3 is there to test.

Literally 1 pixel difference.

Nice.

New Main Posts

0

Q: why use flask session to handle session management failed?

I tried to use flask session http://flask.pocoo.org/docs/0.10/quickstart/#sessions to handle the session management. It turns out sometimes it got kicked out when accessing a protected page, and I have no idea why. any ideas what went wrong here? Thanks!! @app.route("/login", methods=['GET',...

python

ATaco

AFK a tad.

Pavel

So, if a value isn't prepended by i, then it just writes to the current cell?

Nathan Merrill

1:09 AM

Question: if you know that a puzzle only has 1 solution, is using that knowledge to solve it cheating?

Pavel

No, I wouldn't say so

El'endia Starman

@NathanMerrill Not necessarily. For instance, Su Do Ku puzzles in newspapers and books are guaranteed to have a unique solution, and that information can help you find it.

ATaco

I have returned.

Nathan Merrill

Right, but if there really is a single solution, then you should be able to solve it without that knowledge

ATaco

If you can prove that only one solution exists than you, by extension, must be able to solve it without that knowledge.

Nathan Merrill

1:15 AM

The knowledge that there is only one solution doesn't have to come from solving, like El'endia said

Pavel

@ATaco maybe L and R read from the left and right tape, and l and r shift the instruction pointer?

ATaco

Instruction pointer or Memory pointer?

Because the Instruction Pointer is shared among the codes, and can only be modified by basic Flow of Control.

Pavel

1:28 AM

Mp, sorry.

ATaco

It occurs to me this can pretty easily be transpiled from brainfuck.

I have made a dreadful mistake

Pavel

Not if you add more intresting builtins so it's actually useable.

ATaco

That's the intention.

It's not directly transpilable, so that's nice.

Pavel

What happens if a thread tries to read from a cell at the same time another one writes to it?

Destructible Watermelon

@ATaco transpiled to, don't you mean?

ATaco

1:36 AM

From.

Also, writes always happen after reads, to prevent that issue.

Destructible Watermelon

so, BF code can be transpiled to your language?

isn't that true of most languages?

ATaco

Yes.

It can be transpiled more easily than most.

Destructible Watermelon

define more easily

Pavel

In fact, it would probably just be the BF code verbatim, except output has to be on another tape to actually read it. Yes?

ATaco

The code:

Destructible Watermelon

1:38 AM

also adding more features won't make that go away

ATaco

1
  r r r
 + + +

Is the same as +++

I'm not entirely concerned.

Destructible Watermelon

why is this image on my computer

Pavel

Doesn't that just repeatedly add 0?

Destructible Watermelon

ATaco

the 1 sets up the first memory tape with a 1, so the + will add 1 and the value in the second tape, then the rcopies that value from the third back to the second.

Destructible Watermelon

1:40 AM