00:00 - 19:0019:00 - 00:00

12:00 AM
@Mr.Xcoder it is :)
I actually got the idea while writing random stuff on Mathematica

@JungHwanMin My solution is like 16 bytes or so
And I usually don't golf Mathematica. It is probably not optimal at all though.

The idea I had was 16 bytes
Probably the same one :)

I just hope I am awake when you post the challenge...
Because I will likely not be if you post it soon :(...

I'm posting it now, actually

That would be great since it's 3AM here and my brain cannot withstand any delay :)
@JungHwanMin Is that your solution too?

12:08 AM
2

Challenge Given a positive integer \$n\$, output the \$n\$-dimensional pyramidal list. Example \$n = 1\$: Objects arranged in a 1D pyramid (line) with side length 1 is just by itself. So, the output is {1}. \$n = 2\$: Objects arranged in a 2D pyramid (a triangle) with side length 2...

@Mr.Xcoder It is indeed

@NewMainPosts Oh boy, you weren't so fast in centuries!

Ah, it was a dupe huh
I thought my idea was great :(

:| I liked the challenge. Sad that it is a dupe.

I waited 10 days to post it :/

12:18 AM
lol I just noticed that you answered the other challenge

sorry guys, but you can try answering the original instead, and I'm pretty sure the answer isn't that different in this particular case

@Mr.Xcoder Perhaps I subconsciously remembered the other challenge :o

it happens; on the plus side, your challenge may become a very good signpost, which is very important, since it's described in a very different way

12:31 AM
@EriktheOutgolfer I don't know how else to link to it...

at least it wasn't me :P
so, there are just a few hundreds of users remaining to choose from ;)
some of which you know

12:44 AM
i asked someone on twitter who was looking for water if they ever found any and they never replied back
is this how the "Gulf" wars happened

Hey, @WW's profile pic is a cat that isn't disturbing to look at!

uh
maybe were not looking at the same cat here
but this one is extremely angry

I meant that it doesn't look immediately disturbing, such as the last profile picture. Anyways, Disturbing ≠ Angry

1:10 AM
Are there any musical people here that can think of some vibrate pattern that would be resemblant of PPCG/Axtell

None/Empty Pattern: it's the shortest possible pattern; we're code golfers after all :p

What is a "vibrate pattern"

0

Advanced Transposition code-golf array-manipulation Inspired by the obscure 3rd optional argument of Mathematica's Flatten function Introduction Matrix transposition involves swapping the two indices in a rectangular 2d matrix. For example, $$\begin{bmatrix} 1 & 2 & 3 & 4\\ 5 & 6 & 7 & 8\\ 9 Ooh, just at the right timing Is there a nice way to notate a 3d rectangular matrix? \a_{i,j,k}\ ? 1:13 AM @NewSandboxedPosts I suspect APL will win for sure. How would you write {{{1, 2}, {3, 4}}, {{5, 6}, {7, 8}}, {{9, 10}, {11, 12}}} in TeX, for instance? Uh, "notate" like that. Oh shoot, jagged arrays... dang, no APL there then. That's not jagged. In the challenge, the submission has to handle ragged arrays too 1:15 AM Why did you specify "rectangular" theN? I'm writing an example involving a 3d rectangular matrix It would be nice if there is a clear way to visualize it using TeX. $$
\begin{array}{cc|cc}
1 & 2 & 3 & 4 \\
4 & 5 & 6 & 7
\end{array}

Probably...

That would work. Hmm.

I have a feeling Adám is going to pull some fancy trickery with Dyalog whenever that challenge is posted.

Althogh...
I'm not sure if it would be clearer than the code format.
For J: Although it has that "permutation transpose", it can't handle ragged array that easily.

1:24 AM
Pro Tip: If you type python code into a node.js REPL it will not work
definetly did not confuse myself for half an hour
@feersum when you get a notification your phone vibrates, I can customize that for axtell

About the new notation: I can't understand which dimension is which.

@Downgoat What are the parameters?

@user202729 I was wondering that too. Do you think it would be clearer if I make it into a 2D matrix, in which each element is a 1D vector?

That's what Mathematica does. Probably, yes.

@feersum how long to vibrate and how long not to: developer.mozilla.org/en-US/docs/Web/API/…
that said, we also need a notification sound

1:34 AM
@user202729 is it better now?
Also, why is there a scroll bar for each mathjax expression...

@JungHwanMin New design → bug.
(pinned message on starboard) →

@Downgoat holy shit you can do that
When were those APIs added
Time to make my wsbsite extra fun on mobile

for APNS
looking at whatwg commit history they dropped sound for web :(
apns ftw :P

I didn't mean for push notifications specifically
Just the Navigator.vibrate method
Never seen a website use it

1:50 AM
in python is not a < b < c the same as (not a) < b < c or not (a < b < c)

Pretty sure the latter
Awh, mobile Safari doesn't support it
come on

@Downgoat it is indeed the latter: tio.run/##K6gsycjPM/7/v6AoM69EIy@/RMFAwUbBCIiNNf//BwA

2:33 AM
Apparently they are 5 different versions of the Web Push API and so far it looks like google and firefox are looking at two different RFCs T_T

laughing at all these tiny baby mobile apps that are just ripoffs of features of craigslist

2:47 AM
someone please tell me why I can't use the same quote inside an f-string expression python parser can't be that stupid

3:03 AM
0

i don't know if it can be made shorter than this def bubble_sort(num): for j in range((len(num)-1)): for i in range((len(num)-1)): if(num[i]>num[i+1]): num[i] , num[i+1] = num[i+1],num[i] return num print (bubble_sort(list(input("enter the array/list elements:"))))

I was just using a website that during page load, would display a bunch of unfilled templates like {{foo}}, that were later replaced when it finished loading.
How is it even possible to make something as stupid as that?

@feersum because they are using Vue or Angular

Well that just moves the question up a level

Angular 2 is so damn slow
I don't know how Google uses it in prod

@feersum SE with MathJax is basically like that...

3:22 AM
@user202729 Displaying the formulas before hte image loads is arguably good, graceful degradation etc.
Displaying templates looks completely idiotic.
So why would someone use a framework that makes the site look totally amateurish like that?

Probably the framework assumes everyone has super fast machine.

It's like web development literally rots brains.

Ok so I implemented Googles Push RFC and Firefox accepts it but Google FCM rejects the format wtaf
@feersum because learning how to properly render webpages is apparently too hard.

Have you guys heard of WebUSB?

reminds me of the guy who did RS232 over headphone jack with WebAudio lol

3:53 AM
My new favorite discovery on github is that @ncasenmare refuses to use Math.pi in any projects. 😂 τ > π

4:04 AM
Well, of course \$tau>pi\$... but does that mean we should use \$pi*4\$? Does that mean we need to use it? (The answer is no)

was more sharing because of the who

is "sharing" an adjective now?

Jul 27 at 16:16, by quartata
Jul 17 at 20:39, by quartata
listen, youre going to have a bad time if you try to be an English major on my feed here
everytime you people do this to me the onebox gets a level deepe

*deeper

you just activated my trap card
deepe is actually a word

4:17 AM
is "deepe" a comparative now?

go to jail
walk dont run

I have bread in the oven thogh.

oh you can wait for it to rise then
im not an animal

Ok thanks
Can't leave the oven on while I'm awaiting trial.

what kind of bread

4:20 AM
The "I ran out of food other than flour so I just mixed things and hope for the best" kind

@LeakyNun It can be used like one, yes.

this is how it was invented im pretty sure
some idiot spilled arm and hammer in the flower

Bane?

It's also how sourdogh was invented. Milk spoiled but times wes desparate, so they made bread anyway.
Most breads were made this way in fact.

4:24 AM
and "more" can be followed by a noun (phrase) as well, "Oh great, more English Majors on my feed"

a wise woman once said that baking is just chemistry for hungry people but that;s wrong. chemistry was invented to try and cover for the all the fuckups that lead to baking
3
thats why its so arbitrary and not-Physics-y

Cookins is an art, but baking is a science, that's why they don't give out nobels for Cooking.

I dont think dynamite would be useful for cooking anyways

Mhm, I have dynamite. I should have put it in the bread.
You never know.

i think theres, an episode of ""Spongebob) like that

4:28 AM
Did it work out well?

i think squidward got sent to gulag

oof. I'm already going to jail so I'd be good not to press my luck any further

minimum sentence here for West Americano Butchery in the Second Degree is only 30 minutes dont be a baby
its just to fill our lockup quota

Oh in that case I could have served it and been back just in time to enjoy my tasty bread

oh oops sorry
checkin times can be long though

4:32 AM
Don't I have a right to a speedie trial.

no
lol

Hm maybe that was something I saw on tv.

you have no idea how often your "TV" has lied to me
countleas inaccuracies wasting hundreds of thousands of hours of Time

You must be confusing me with someone else. i don't own a tv.

just a month ago i found out that you dont actually have robots that can drink beer or smoke cigars yet
@WW sorry it was a Royal You

4:36 AM
My downstairs neigbors do thogh. And sometimes I pickup misconceptions throgh the floorboards.

sue them

It's mostly the landlords fault, he's supposed to fix the leak
I may have added too much yeast, the bread has expanded out of it's container and continues to grow at an alarming rate.

oh I didn't know it was a yeast bread
let me know if you need some of the flamethrowers like in stranger things

Yes, surprisingly I did have yeast
That was a yes to the yeast part not the flamethroughers.
I don't want to burn the bread

5:02 AM
*flamethrowers

5:31 AM
@quartata This reminds that I still have to remove Pi from Cheddar & VSL and replace them with tau...

5:47 AM
@Downgoat ... don't

@Zacharý Pi is wrong and I will not stand for it

Just give a deprecated message.

@Zacharý wrongness doesn’t deserve to be deprecated

Even better: create a sheep-error.

6:07 AM
@Zacharý I have a question about RAD
is it intentional that ⍉2 3 4⍴1 returns a 3x8 array (i.e. list of lists)?

@ngn Go ahead and shhot.
@ngn No, that's not supposed to happen. I'll look at it later.

@Zacharý I'm trying to think about what's a good definition of "transpose" (⍉) in k's model (only lists, no solid multidimensional arrays). In particular, what ⍉1 and ⍉1 1 should return.

1≡⍉1, 1 1≡⍉1 1.

@Zacharý you return ⊂1 1 for ⍉1 1
which seems more correct to me :)

Yeah, that does seem more correct.
@ngn Maybe look at the sandbox question posted above somewhere?

6:14 AM
@Zacharý I would argue that ⍉1 (transpose of a scalar) should be a 1x1 matrix

@ngn Well, the reason 1≡⍉1 in RAD is to align with APL, since scalars (without boxing of arrays, obviously) shouldn't cause rank issues.

@Zacharý but in APL it makes sense - as you have multidim arrays, "transpose" reverses the axes
for a scalar the list of axes is empty, so it's empty when reversed, so it returns the same scalar
but in k and RAD, we're only swapping the leading 2 axes
and this becomes a problem when there are fewer than 2 - in scalars and vectors

@ngn Transpose is implemented as some weird stuff in RAD: ⍉ is ,¨/ right now. That needs to be fixed.

@Zacharý I agree. That would work on matrices consisting of scalars, but would fail on many other kinds of arrays.

It could be ⍴,⍤0 0 0/, but that gives 3 4 2≡⍴,⍤0 0 0/2 3 4⍴1.
Which actually ... might just give me a one-byte solution to the sandbox challenge if I make a certain ⎕MISMATCH default. LOL.

6:24 AM
@Zacharý wouldn't it be easier to implement ⍉ with its own D code, not through other primitives...

@ngn Yeah, probably. But ⍺⍉⍵ is going to be a MAJOR pain to write (and even think about, having ragged arrays everywhere)
Code like 0,2 2⍴⍳4 is vastly different from APL, so I'll probably have to add another atom somewhere.

@Zacharý yeah, in k's model permuting the axes would probably be hard to implement, though in apl's model it's surprisingly easy

Does ⍪⍵ => ,⍤0 0 0⊢⍵ and ⍺⍪⍵ => ⍺{⍺,⊂⍵}⍤1 1 1⊢⍵ sound close enough to APL? I'm redoing ⍪, since I generalized strokes too much
I could just check ⌷ before doing the permutation

@Zacharý what does the thrid 0 in ⍤0 0 0 mean?

@ngn ⍤ is MONADIC LEFT RIGHT. I need to fix that as well

6:36 AM
@Zacharý in Dyalog ⍪⍵ leaves the leading dimension as-is and merges the rest
shape of ⍵    shape of ⍪⍵
a b c d       a (b×c×d)
a b c         a (b×c)
a b           a b
a             a 1
empty         1 1

So ... ∊¨ in RAD (except for ⍬)?

@Zacharý yeah, I guess so
note the implicit trailing 1s in the shape; if we consider those for ⍉⍵, then ⍉1 naturally becomes a 1x1 matrix
"shape" of x    "shape" of the transpose of x
in APL          in k
a b c d         d c b a         b a c d
a b c           c b a           b a c
a b             b a             b a
a               a               1 a
empty           empty           1 1

⍴⍪⍬ is giving 0 1 in Dyalog...

@Zacharý well, makes sense. look at the second to last row in my table for ⍪

Yeah.

6:49 AM
⍬ has shape ,0 so a=0 and the shape of ⍪⍵ is 0 1

Oh, got ⍴ confused with ≢, never mind.

I tested kona and oK - they both return the scalar as-is when transposed
I think they're both wrong :)

1 hour later…
7:52 AM
19

I'm sure most, if not all, of you have come across this at some point or another: Aoccdrnig to a rscheearch at Cmabrigde Uinervtisy, it deosn't mttaer in waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht the frist and lsat ltteer be at the rghit pclae. The rset can be...

holy shit wtf google
so I have spent the last few hours on why google won't accept a JWT
if the JWT is encoded with spaces between the keys google rejects it

8:08 AM
.oO Wai... Wut? <insert look of disapproval, Google!>

5 hours later…
12:57 PM
@ngn I wonder what a transposed scalar is...

@EriktheOutgolfer see the discussion above and this bug report

Come on? Is anyone going to post an answer in this month's LotM?

it's particularly inactive

Actually... SMBF is just BF + short constant representation. It is not very practical to modify the code.
There isn't many challenges which has both constants and simple algorithm.

@ngn so, you're suggesting that, in APL terms, (1 1⍴1)≡⍉1 and (1 2⍴1)≡⍉1 1?

1:06 PM
@EriktheOutgolfer yeah, I'm suggesting that only for k

oK or general?

@EriktheOutgolfer I don't think Kona developers would be eager to change this, they tend to stick to an old version of K. Also, theirs is slighly "less wrong". One could argue that they reverse the leading min(2,rank(x)) dimensions.

@ngn does oK throw an error?

anyway I'll notify Kona a bit later
@EriktheOutgolfer no, it returns the scalar as-is

@ngn it looks like it encloses vectors though (or at least makes them into a matrix), right?
yeah, if that's the case for vectors, I'm sure that scalars should be handled as singleton vectors (or singleton matrices)

1:13 PM
@EriktheOutgolfer yup, that's why I claim oK is wronger, it's inconsistent
@EriktheOutgolfer hm... why?

@ngn +1,,1 makes sense, no? (note: not a K expert)

by making it a singleton vector or matrix, you're prepending 1s to its shape, whereas I would consider implicit trailing 1s in the shape to be more natural

but a scalar's shape is the empty vector, no?

@EriktheOutgolfer yes

so it should just be made into 1 1, [] + [1, 1] is the same as [1, 1] + []
(+ = concatenation, Python terms)

1:17 PM
the question is how do we swap the first two elements of an empty vector :)

well, a scalar is a value, "a value" → "one value" → "1 value"

@EriktheOutgolfer what about ⍉1 1 then?

@ngn doesn't that return (⊂1 1) already? or is it (1 2⍴1 1)?
(the two are kinda different, that's why I'm asking)

@EriktheOutgolfer it does (more like ,⊂1 1 in APL or ,1 1 in k), but how do you justify that if you're prepending 1s to the shape?
@EriktheOutgolfer in k (and Zachary's RAD) they're the same

@ngn if a vector's shape is 2, then the shape of the respective matrix should be 1 2, no?
@ngn yeah, k doesn't have ranks

1:21 PM
@EriktheOutgolfer yes. "shape" of course being an informal way to talk about rectangularly nested lists

@ngn yeah ik, but, if you append ones instead of prepending them, it would become 2 1, which will result in [[1],[1]]

@EriktheOutgolfer remember - with +x (k's transpose) we're swapping the first two dimensions
so if a vector x has implicit trailing 1s in the shape, one of those 1s will come before the vector's (original) length

@ngn if we consider ragged arrays, I'm pretty sure that [1,1] should become [[1],[1]] by itself, before transposing, as [1,2,[1]] should become [[1],[2],[1]], no?

@EriktheOutgolfer I don't know how ragged arrays should behave under transposition... I haven't thought much about it.
I only know that k does something like scalar extension:
 +(1 2 3;4 5 6;7)
(1 4 7
2 5 7
3 6 7)
the above works in <s>all</s> three free(ly accessible) dialects of k - kona, ok, and mine

 +(1 2 3;4 5 6;7 8)
(1 4 7
2 5 8
3 6 7)
I'd expect something like this to be the case too
@ngn who knows if nobody has written a fourth one? ;-)

1:30 PM
@EriktheOutgolfer right... :)
I've just tried to strike out the word "all"... never mind
@EriktheOutgolfer this doesn't work in all version of k

@ngn ---
(too late now)

In k5 that used to be an error. In ngn/k I decided to have it as an error too.

it looks like it does that in oK

@EriktheOutgolfer I don't know if that's deliberate... the manual mentions only atom spreading (i.e. scalar extension)

@ngn well, that's not the only possible behavior
there's the error too :P

1:40 PM
@EriktheOutgolfer right, I was also thinking of zero-filling, like APL's "mix" (↑⍵), instead of scalar extension

hm, maybe that's a bit too much

+(1 2 3;4) -> (1 4;2 0;3 0) instead of (1 4;2 4;3 4)
or maybe even 0N-filling (with int nulls)...
anyway, at this stage it's better to be as compatible with the others as reasonably possible

2 hours later…
3:40 PM

3:56 PM
hi all
Given a binary string X of length 100, say, how hard is it to produce another binary string with edit distance k from X?
On a related note.. just posted a new sandbox edit distance related challenge
hopefully people will enjoy it

Just append k 1's at the begin.

@user202729 good point!
I meant of the same length :)
maybe that makes a decent ppcg challenge too.

Any restriction on k?

not for this question no

0

Edit distance for sparse strings The input to this challenge will be two strings of length one million each. Each string contains only zeros and ones. However each string will contain at most 100 ones and so will be represented by a sorted list of integers. The integers will indicate where the ...

4:02 PM
Then it may be impossible to construct any such string.

@user202729 sure.. in that case no output is valid

I was in the bus today, and noticed this sign:
      seated    standing    wheelchairs
max1    37         30            00
max2    36         26            01
max3    34         32            00

What is max in this situation ...?

I wonder what that could mean?

How does a challenge sound where input is such a matrix and a list of [seated,standing,wheelchairs]. Objective is to determine if the bus load is allowed.
It means there can never me more than 37 seated passengers, and only if there are no more than 30 standing and no wheelchairs.
There may be a maximum of 1 wheelchair, but then there may only be 36 seated and 26 standing passengers.

4:06 PM
... understood. So it's allowed iff one of the row has all entries greater than the second input.

There may be up to 32 standees, but only if there are no more than 34 seated passengers and no wheelchairs.
@user202729 Exactly.
In fact, there could be any number of rows.

And we could even generalise it to have any number of columns too (with the other input matching in length to the columns).

And linear interpolation?
(for example, given
20 0
0 10
, 10 5 should be allowed. Not in practice)

@user202729 No, the mathematician in me wants to calculate the seated vs standing vs wheelchaired equivalences, but I counted the seats and realised why those limits where given.
@user202729 So, 10 0 and 0 7 would be ok, but not a mixture.

4:13 PM
I still don't get how that bus works.
And... I guess allowing linear interpolation would make this a linear programming problem.

@user202729 one seat folds away to grant space for a wheelchair. Two seats fold away to allow people to stand.
Besides for blocking a folding seat, a wheelchair also uses the floor space of four standees.
The area gained by folding away the seat in the wheelchair area is not safe to stand in.
Each one of the regular folding seats can be folded and one person can stand in its place.
@user202729 Yeah, with a bunch of boring built-in answers.

4:48 PM
cs researcher: we need to figure out ways to write safer code with fewer bugs so it can be exploited less often. Hu et. al.: what if *takes a huge bong rip* we added more bugs to the system instead. https://arxiv.org/pdf/1808.00659.pdf (this paper is lit)
3

-4

Good evening, dear forum users. For educational purposes, I wrote one software. But, at startup, I see the error "The application is in the suspend mode" in Visual Studio. I do not have syntactic errors. What could be the problem? I'll leave the code below. using System; using System.IO; usi...

5:00 PM
do people think this is better as is as a code-golf question or as a fastest-code question ? codegolf.meta.stackexchange.com/a/16753/9207

I don't think it's a good fastest code

@WW Thanks. Could you say more about why?

I also don't think it makes a cood code-golf as is

oh.. can you say more about that too in that case please

tbh the codegolf solutions are all basically going to be fast anyway

5:04 PM
that's not 100% clear to me

What you have pegged as the naive algorithm seems much slower to me that the algorithm I would reach for.

@WW oh what would you reach for?

The naive alogithm would just be to count the number of items that are in exactly one list.

@WW if you do flips only it is always count the differences. But you need to get the optimum which can be different due to shifting also being possible

My algorithm as a haskeller would be slightly different, but ultimately the same.
@fəˈnɛtɪk Shifting?

5:07 PM
add a zero in front remove one behind to shift a bunch of them

Ah ok

ie 1,3,5 to 2,4,6 would be 6 for flips only but 2 for shifting

Ok the challenge still inst very good as either fastest code or code-golf with time restriction.

How do I downvote a website theme?
37

tsk

5:20 PM
2

I was in the bus today, and noticed this sign: seated standing wheelchairs max1 37 30 00 max2 36 26 01 max3 34 32 00 The number of seated passengers, standees, and wheelchairs all have to be no larger than some row i...

@PhiNotPi downvote the meta post? ;)

I think th e emerging method is to star phinotphi's chat message.

sure
that's how we do pleas here ;)
@JonathanAllan dupe comment :P
[30,31,0] → [0,0,1] I think — Jonathan Allan 17 secs ago

5:45 PM
@WW That method would often give the wrong answer

6:08 PM
For some reason, I can't get ⎕IO to work...

Yeah.
I can change it manually through the D code, but changing it via RAD isn't working
It's parsing it as ⎕ IO for some reason

I haven't implemented any other quad-names. Just ⎕IO for now (I should probably implement ⎕UCS though, considering I just make it do the same as " does)
Fixed it: I need to put the multichar atoms before the single-char ones.
There. ⎕UCS has been added, and multichar-atoms are also fixed. Just waiting for a pull
I broke negatives....

6:29 PM
@Zacharý :-(

That's probably for the best due to floating point errors. More general than just 0, and more so than 0/1.

@Zacharý Wat?

I have no dang clue, but that could be useful.

@Zacharý You have ⍳⍬ give ⍬. It should be ⊂⍬

@Zacharý Hmm

6:34 PM
Uh ... I have a feeling floating point is broken now. That explains everything going on with ⍳  (except ⍳⍬)

@Zacharý I still get ⍬ for ⍳ on negative ints after Dennis's pull.

I mean: I broke the PARSING of negative numbers. ¯1 was being parsed as ¯ 1 (¯ will be negative infinity)

@Zacharý Right. I noticed.

Right now, I'm going to fix floating point. I knew I was going to break something in the speed-up

@Zacharý Are lowercase names prohibited?

6:38 PM
@Adám For nilads, yes. f←+ should work though. First letter uppercase: nilad; lowercase: funciton; ∆: monadic operator; ⍙: dyadic operator.

@Zacharý Shouldn't 2+3i×3 give 6+9i? Try it online!

Is it not just right-to-left?

@Adám No, that's intended. Notice how complex numbers are outputted in parentheses? That's because there are only imaginary literals, not complex. So 2+3i×3 is like 2+0J3×3

@Zacharý If so, then RAD has no complex input format?

@Adám Nope. I didn't want to have to worry about complex-numbers colliding with base-notation.

Lol.

0

Work out Naismith's Rule code-golfnumber Naismith's rule helps to work out the length of time needed for a walk or hike, given the distance and ascent. Given a non-empty list of the altitude at points evenly spaced along a path and the total distance of that path in metres, you should calculate...

@LeakyNun that makes me wonder, does 0 exist?

0 !== "0"
Wait ... "\t" == 0?!
Holy crap, that's a good test for whitespace ... if it wasn't for "0", "0.0", etc.

@Zacharý if one of them is a number, js's == tries to convert them both to numbers

00:00 - 19:0019:00 - 00:00