JPEG XL

btw <@794205442175402004> are you able to tell me any curiosity of the process process or the format itself 😄 like that our professor told us we will get a 10 hahaha

I mean a not well known fact, he said us that if we surprise him he will increase our mark

I'm curious - did the professor choose a bunch of graphics formats and assigned different one to each group? FLIF itself is a rather niche format those days.

we had to choose from a wide variety of video/audio/ image formats and we decided to choose this one

this was the list

Apple Video, AV1, AVS, Bink, Cinepak, Daala, Dirac, DV, DVI, FFV1, Huffyuv, Indeo, Lagarith, Microsoft Video 1, MSU Lossless, OMS Video, Pixlet, ProRes 422, ProRes 4444, QuickTime (Animation, Graphics), RealVideo, RTVideo, SheerVideo, Smacker, Sorenson Video, Spark, Theora, Thor, VP3, VP6, VP7, VP8, VP9, WMV, XEB, YULS, ACELP, AC-3, AC-4, ALAC, Asao, ATRAC, CELT, Codec2, DRA, DTS, FLAC, iSAC, Monkey's Audio, TTA (True Audio), MT9, Musepack, OptimFROG, OSQ, QCELP, RCELP, RealAudio, RTAudio, SD2, SHN, SILK, Siren, SMV, Speex, SVOPC, TwinVQ, VMR-WB, Vorbis, VSELP, WavPack, WMA, MQA, aptX APNG, BPG, DjVu, EXR, FLIF, ICER, MNG, PGF, QTVR, WBMP, WebP 3GP and 3G2, AMV, ASF, AIFF, AVI, AU, BPG, Bink (Smacker), BMP, DivX Media Format, EVO, Flash Video, GXF, IFF, M2TS, Matroska (WebM), MXF, Ogg, QuickTime File Format, RatDVD, RealMedia, RIFF (WAV), MOD and TOD, VOB, IFO and BUP

Lord help him when he has to mark all the obscure ones haha

These were the 7 proposals: - PIK - FUIF - AVIF (later retracted) - TAT - DIC - XAVS2 - something I don't remember / cannot find the name of, based on lapped transforms

about FLIF or about JXL?

FLIF please

FLIF was originally called JPIF, for Jon & Pieter's Image Format, though Pieter insisted it should be called JIF to maximize confusion with GIF

wow thank you !

it’s a shame that JXL wasn’t available to us for choose

Could've gotten bonus points if he hadn't heard of it before ;P

Thanks. Interesting to know AVIF was among the proposals.

No. For me 6 cores is more or less enough. But I was thinking about some upgrade, because things like av1 are quite computation-heavy.

I find this list a little odd, since many of these are very well documented by international sttandard, some are obscure manufacturer-related, some are containers, and some are codecs, some for image, some for video, some for audio

there's a lot of choice here, so you could probably have asked about JXL

for example, 3GP and 3G2 are just variants on ISOBMFF, which itself is derived from MOV

but there's really not much to talk about ISOBMFF

beyond the box structure and required boxes

yeah i think it was on purpose to make us investigate on what to choose

Oh my god its such a big die

So, I’ve been feeling discontent with the performance of my chess bot’s library API, so I decided to investigate improving it. As of now, I keep two chess implementations on the bot’s repository, one that is nice to use, for the library API, and one that is fast, to use during evaluation search. I came up with a different implementation that is faster than the current one. As a bonus, it allows two forms of usage: an immutable API and a mutable API. I have decided to benchmark it, and these are the results I came up with: ``` 12002.8765ms / 4894325 = 0.0025ms (deepen) 12258.6647ms / 4894325 = 0.0025ms (getMoves) 14885.8047ms / 120909581 = 0.0001ms (fast-chess.js) 2738.1347ms / 197742 = 0.0138ms (standard) 18466.8493ms / 197281 = 0.0936ms (old) ``` - The first value is the time it took to complete the benchmark for the given implementation. - The second value is the number of positions traversed. - The third value is the average time per position. - Each implementation has special moves (castling & e.p.) disabled. - In each implementation, the king can be put into check and even be captured. The meanings of the names in those benchmarks are as follows: - “deepen” and “`getMoves`” are names of slightly different usages of the new mutable API. - “`fast-chess.js`” is the name of the existing mutable implementation used during search. - “standard” is the name of the new immutable API. - “old” is the name of the existing mutable API (exposed for the library). Now, the new API is faster than the existing API, which is fantastic! Though it doesn’t seem to be able to beat the existing `fast-chess.js` implementation at all! (Even though it is based on it.) This is good overall, because in addition to this “no special moves, kings can be captured” variant, the new API also implements standard chess, and also allows me to encode more variants, which was one of the goals I had established when working on it. The fact it is faster than the current API implementation means I can replace it with this new one without a performance penalty. (Those who saw my last message about adding variant support to my bot’s library will remember it was very slow, and also very complicated, which this new implementation solves completely.) Now, this is not the most ideal, because I was hoping to also replace the `fast-chess.js` implementation with this new one (i.e. unify the implementations altogether), and I can’t seem to be able to currently, unfortunately. So yeah, I’m not sure whether I should feel happy or disappointed by these conclusions. 🤔

https://twitter.com/UniverseIce/status/1648619132760256512?t=9IeSBpoYOX0m7ZZmq6UPJw&s=19

This is done on purpose.

Wow, good on them

Is this a comparison between two terrible cameras or a good one and a bad one?

Two smartphone cameras, likely using very similar sensors

smartphone photography is very much always a battle between two evils, with a decision on what's "better" being made depending on which is the lesser

I have a battery which works opposite 🤔

The undersharpened one is nice in its own right, but not great in general. The oversharpened one is garbage tho.

Lawsuit!

How does it feel?

Does gamma basically change the range?

I don't notice anything

Nice

But this isn't a laptop. It's an all in on pc from Lenovo. It doesn't run on batteries.

No

Exactly

"hey, we can make your porn more efficient rather than just deleting it all"

Problem is, bandwidth is also a concern along with storage. They wouldn't be able to serve jxls since most browsers don't support it.

sound like an expensive plan to push forward jxl adoption - create jxl-only imgur archive, some ppl would install jxl extension for this 🤪

But most would probably just stop visiting or using their site.

I mean it doesn't sound like the worst idea ever since the images that imgur is removing are images that not many people access and also imgur wants to get rid of them for saving space

the problem isnt bandwidth or storage

it's just investors

so much of the traffic is for reddit and more people will simply not use it since they cant

Why PDF readera are still rendered in software?

Does any hardware pdf renderer exist? That would be extremely surprising to me...

That's a philosophical question about a nature of printers 🙃

I would assume printers use some internal cpu to rasterize stuff, if it isn't done by the printer driver already

Ancient printers probably handled postscript commands by actually doing in hardware what the commands say, but I don't think anyone would still dare to do such a thing nowadays

you know what i dislike about printers, or at least the one i have? it scans papers to jpeg, but no software i have seems to be able to get the original jpeg from the printer, so i either have generation loss if i save it as a jpeg again seemingly, or i save it as a bigger png

however i recently noticed that even with the extra jpeg generation loss, jpeg quantsmooth can still restore it pretty decently, which i wish i found out before i scanned a ton of my old drawings to png and converted them to webp (which im considering converting all to lossy jxl)

Are you sure it scans to jpeg? Is that what goes through the wire, or is it what the printer driver produces or the scan app?

i have tried about 3 different apps and it all looks the same with the jpeg artifacts

normal windows fax and scan, NAPS2, and VueScan

this is what every scanner app gives me

and this is what i can get from jpeg quantsmooth, so it's a little better

also it's a wireless printer

you also tried the manufacturer provided program as well from what I remember

which also resulted in either the same or a worse result

that didn't even listen to DPI resolution changes lol

I tried to find out if the JPEGs that programs where giving you where native or not but i'm not enough of an expert to say whether or not they have generation loss

there is some slight artifacts around the pencil however still

I think what programs are giving you might be native JPEGs

its a 600 DPI scan, and a little more noticeable in 1200 DPI

they are just forced 4:2:0 as that's what the printer probably natively outputs

I would hope at least

either way it seems like either there was no generation loss or the generation loss only a very small amount since quantsmooth and the like seem to get rid of all the blocking artifacts

<@146411656174501888> if you take this image for example you can tell it's been JPEG'ed multiple times because stuff like quantsmooth and jpeg2png are unable to remove the artifacts https://i.imgur.com/GREQj.jpg

Yeah, pure hardware pdf renderer could fit nicely into some steampunk universe.

There was a paper about the rendering being moved to GPU https://web.engr.oregonstate.edu/~mjb/cs519/Projects/Papers/AcceleratingVectorGraphicsRendering.pdf

<@146411656174501888> did you try scan to PDF and then extracting the embedded jpeg in the pdf?

only with NAPS2, and it was a lower quality jpeg (quality 80 i think?), i still need to try that in other programs

Almost all the printers/scanner I ever used have the option to export pdf, jpg and tiff.

just a simple printer/scanner, HP Envy 5660, and yeah, i should try it with linux, i didn't even need to install drivers to get it to work with windows, but idk how my linux mint install will handle a wifi printer

I got it setup pretty easily when I did it

That was mint in 2018, 2020

i'll have to test that out on my laptop later, which is iffy as the battery is basically dead on it, and where the printer is, i can't connect it to usb directly to my pc, and i can't move the printer anywhere

Haha lol

I wish. I would love to do that, but ISO doesn't have that kind of sense of humor, I am afraid.

Isn't that just paper?

I think they meant a hardware chip that rasterizes a PDF for use in printing

but I guess "hardware PDFs" could be paper <:KEKW:643601031040729099>

Solid state PDF storage

Don't think so, but x264 had a tune just for Touhou replays

Anybody watch SAO: Abridged?

too old

Then perhaps you'd enjoy the humor 🤔

I'm still waiting for more episodes lol

<:Thonk:805904896879493180> || no || ||?||

Well the writers for the abridged version have very dry humor with lots of classic movie references that us uhhhh recenter generationals wouldn't get

Considering the size of the PDF specification and the number of edge cases, it's going to be a huge chip

Typically hardware acceleration just does the most common operators (the "hot" loops), and then all the various extra stuff happens in software/firmware.

Take out the weakest links, or rather the heaviest chains

That's a lotta loops.

So it could give the wrong colors at times?

Yeah but they said 'pure' hardware PDF renderer

?

Aren't sums loops?

https://arxiv.org/pdf/1711.03951.pdf I took the formula from here on the second page

the Σ everywhere and lack of parentheses make it look more complicated than it is

α = ((M × N) · sum(L .* C) − sum(L) · sum(C)) / ((M × N) · sum(L²) − sum(L)²)

🙂

isn't this just least squares

does anyone know what this method of placing objects is called? I watched it for 20 seconds and figured out how to do it. it's quite simple. I would like to know the name of the method in order to try ready-made implementations

sphere packing?

thumbnails are still sorta messed up but that's to be expected as it's always been like that for discord

I can't understand latin, but I like this song. 🙂 > https://www.youtube.com/watch?v=CyrsVvea0TM

not at all. not that at all. can use any objects

This is done elementary. We are creating a cube. Placing the object. And make lines the size of an object to each side of the cube. so we get 4 mathematical holograms. and this place will be considered occupied. And the object will be visible from each side of the cube. So do it in hundreds of objects and you will get this kind of animation. It's strange that I can't find anything about this

well https://www.reddit.com/r/woahdude/comments/103zlzh/perspective_packing/

https://thumbs.gfycat.com/RealisticSlightGrassspider-mobile.mp4

now I see that it seems to have no name..... and for some reason it is done only with spheres

https://www.shadertoy.com/view/ctjGRD

as it turned out, literally the only one person on the Internet is doing this

https://en.m.wikipedia.org/wiki/Packing_problems this?

"packing objects perfectly on 2d" leads to some nice search results

Link collection from search results: 2D: <https://www.gorillasun.de/blog/a-simple-solution-for-shape-packing-in-2d/> <https://www.nature.com/articles/srep16722> <https://pubs.acs.org/doi/10.1021/jacs.0c08979> <https://math.stackexchange.com/questions/3148483/how-do-i-pack-2d-objects-into-an-arbitrary-shape> <https://github.com/topics/rectangle-packing> 3D: <https://en.m.wikipedia.org/wiki/Close-packing_of_equal_spheres> <https://pubs.acs.org/doi/10.1021/acs.iecr.8b03915>

But yeah as for the animation... that's nice but will be hard to find anything like that

no

<@230800661837512705> This is quite literally nothing to do with packaging at all

It's not packaging at all. This is just a way to create a beautiful illusion

math illusion

your links literally say "perspective packing"... this is a packing problem, just a different kind (not aiming for density, but perspective)

This is the name given by this person. But it is not true. I followed his work and previously he was engaged in Packing spheres inside a 2d frame. But the 3D version that is being demonstrated is not packaging. Algometric it works so that it only generates spheres in free areas -- There is no solution to the problem of packaging

I would call this a perspective placement of spheres

I wrote an algorithm on paper. which would work. but unfortunately I am not a blender programmer to implement this. BUT this is done very simply

take any point. draw lines perpendicular to each side of the cube... That is all. this space is occupied

I also found out that with such an algorithm, it is impossible to intersect spheres. as well as the fact that the amount of occupied space on each side of the cube will be completely identical

mostly nothing. besides the fact that it looks interesting..... as well as the fact that it can be done for absolutely any geometric shape. as well as the fact that the placed objects can be of any shape

perspective placement

Yeah okay but effectively when looking at one side of the cube you get a 2D packing of spheres

In some way it may be considered a packing problem

However yeah, hard to find something like this

Cubes like this could make some nice gifts / things for a key chain 🤔

If we had spheres given to us before creating the cube, we would solve the packing problem. but IT generate spheres wherever it is convenient, without solving anything.

Unfortunately, this animation is based on the fact that the viewing angle = 0 degrees. that is, in reality, such an effect we could not achieve. only if we made a giant monument measuring 100 stories, and would look at it from a distance of 10 km

https://youtu.be/ur560pZKRfg?t=42s

these are the lyrics I sing now

“you shine like salt and this is me”

“far from fleeing and yitewewoteli”

https://open.spotify.com/playlist/6EsPDvSWnWvrwPmLQOZEvc?si=749f31fc8a31419f

Anyone else feel like they make certain observations at a regular interval?

Yes. I always check the time when it is 10:08

I feel like my memory/perception is based in 5 second intervals.

https://youtu.be/E7RT4wNhiYQ

2h24m? TL;DR?

Is this just a promo video for https://en.wikipedia.org/wiki/Christian_Science ?

> There are key differences between Christian Science theology and that of traditional Christianity.[11] In particular, adherents subscribe to a radical form of philosophical idealism, believing that reality is purely spiritual and the material world an illusion.[12] This includes the view that disease is a mental error rather than physical disorder, and that the sick should be treated not by medicine but by a form of prayer that seeks to correct the beliefs responsible for the illusion of ill health.[13][14]

Sounds like it's basically the Church of Placebo

The placebo effect does exist and there is some amount of 'healing by sheer willpower' that actually works. But viruses and bacteria, genetic conditions, nutritional deficiencies, etc also exist 🙂

even in the first 10 minutes, he describes attempting it and failing(taking medicine)

i'm beginning to feel youtube infotainment burnout

actually not really: what people think of as “the” placebo effect is vastly overrated, most of it being just regression to the mean, spontaneous recovery and reporting bias https://www.painscience.com/articles/placebo-power-hype.php https://sciencebasedmedicine.org/placebo-myths-debunked/ https://www.researchgate.net/publication/38415685_Three_things_every_medical_writer_should_know_about_statistics http://www.dcscience.net/2015/12/11/placebo-effects-are-weak-regression-to-the-mean-is-the-main-reason-ineffective-treatments-appear-to-work/ https://onlinelibrary.wiley.com/doi/full/10.1111/j.1365-2796.2004.01355.x

it is true that the body has impressive self-healing capabilities but it’s not going to wait around for a fake pill to kick them into action

it would be surprising from an evolutionary point of view

as https://www.painscience.com/articles/laser-therapy.php puts it: > The hopes of all kinds of regenerative medicine are all pinned on the grand idea that biology has potential for improved healing that can be “unlocked.” But this is broadly implausible. Biology doesn’t really have many valuable hidden features that only emerge with an exotic/rare stimulus: either a system didn’t evolve in the first place, or it did and we’re already using it. > > But biology is also insanely complex, so who knows?

https://cdn.discordapp.com/attachments/803196804651941959/1102532709705859072/7ec961399aba8830.png

I fell for the Mandela effect and made this meme. I always knew that all video cards support VP9 hardware encoding and even asked - why do they do this? - and they answered me, they never support VP9 hardware encoding, - I started checking.... Surprisingly, it is. never in history. Decoding only. I also found hundreds of messages in the AV1 community's discard server where people literally say VP9 hardware... Amazing Mandela effect however.

Can't find any messages over there that claim GPUs support VP9 hardware encoding though...

``there is hardware vp9 encoding in all intel since kabylake`` "Main and Main10/10-bit, and VP9 10-bit and 8-bit video.[21][24][25][26] Hardware encode is supported for H.264 (AVC), HEVC Main10/10-bit, and VP9 8-bit video. VP9 10-bit encode is not supported in hardware." according to Wikipedia ``and everything else is either using EVE-VP9 or a hardware encoder`` No mention of VP9 GPU/CPU encoding ``well afaik the hardware vp9 encoder was so buggy that intel force disables it now :kekw:`` response to a screenshot of Kaby Lake hardware encoding table from Wikipedia (which shows VP9 hardware encoding) Too lazy too look at the others

Most of the posts from those screenshots aren't about CPUs and/or GPUs having hardware encoding

I'd imagine a lot get mixed up with HEVC/H265

I know that's what I thought until I corrected myself after thinking for a minute

Yeah or I guess in this case prayers or something rather than a fake pill, since I assume these folks don't want placebo medication either.

https://xkcd.com/2726/

don’t forget the classic: https://xkcd.com/2530/

Never seen KB, I take it?

It irks me that people call the body's ability to heal impressive as though there's anything to compare it to.

eh, “impressive” is a subjective term, I don’t feel that there is a super strong need to justify it rigorously

Ye ye it's not a big deal. People are allowed to be amazed by *healing*.

I just take everything literally.

It's not a promo. KB does extensive research to try to bring the viewer as close as possible to the perspectives of those being researched. He is kinda known for this and even had an April fools where he shilled ancient aliens.

that's interesting but I don't really have time to watch a 2h24m video very carefully debunking some bogus religion — though it's nice that such a thing exists, in particular for those who are into that bogus religion

I usually play it while debugging lol

christian science became very popular before medical technology took off and nothing really worked well

but now that medical science provably works and faith healing provably doesn't

it's become a lot less popular

it's interesting how the label "science" gets used by these religions (also scientology) — as if they want to imbue their thing with some intellectual authority (without actually adopting any kind of scientific methodology, obviously)

reminds me a bit of how the nazis used the label "socialism"

The only science they have mastered is manipulation.

that's more of an art :)

It could also be called an art, yes.

Absolutely.

(Whew! Almost got me!)

hm?

I am reminded of this excerpt from Jaynes’ book on probability theory:

(“ESP” = extra-sensory perception)

with this additional note: > Note that we can recognize the clear truth of this psychological phenomenon without taking any stand about the truth of the miracle; it is possible that the educated people are wrong. For example, in Laplace’s youth educated persons did not believe in meteorites, but dismissed them as ignorant folklore because they are so rarely observed. For one familiar with the laws of mechanics the notion that ‘stones fall from the sky’ seemed preposterous, while those without any conception of mechanical law saw no difficulty in the idea. But the fall at Laigle in 1803, which left fragments studied by Biot and other French scientists, changed the opinions of the educated – including Laplace himself. In this case, the uneducated, avid for the marvellous, happened to be right: _c’est la vie_.

Is this the same kind of thing as where Chinese alternative medicine was invented in the 1940/50s because post revolution there wasn't enough actual medicine to go around.

(Fraetor is referring to https://slate.com/technology/2013/10/traditional-chinese-medicine-origins-mao-invented-it-but-didnt-believe-in-it.html)

> “Even though I believe we should promote Chinese medicine,” Mao told him, “I personally do not believe in it. I don’t take Chinese medicine.” > > Mao’s support of Chinese medicine was inspired by political necessity. > > […] > > Mao knew such medicine would be unappealing to empirically minded Westerners. He knew this because it was also unappealing to empirically minded Chinese people.

https://discord.com/blog/usernames

> After we transition over the course of this year, everyone will have: > > 1) A unique alphanumeric username without a discriminator, such as “@phibi”. Think of this merely as a technical necessity. This username will be used only for connecting with other users (friending) and telling users apart when you check their profiles. They’ll be limited to lowercase characters (a-z), numbers (0-9) and two special characters (period and underscore). They’ll be much easier to remember, verbalize and share with your friends. You’ll be able to continue to change your username, but not too frequently.

And then a display name that can be anything.

Seems reasonable enough.

Ye I just yanked the relevant info because that article is so annoying to read.

https://blogs.gnome.org/shell-dev/2023/05/04/vivid-colors-in-brno/

Ooh

Shouldn't gnome work on fixing their software before adding new features?

Arguably this is fixing how colours are done.

'bout time

Good point lol

My gripes with gnome: You can't disable notification popups globally, only for individual apps. I still want notifications, just no popups. Literally switched to a modal text editor because I hate popups so much. So... Much... You can't screenshare with discord, tho that might be a compatibility issue. Typing password doesn't automatically focus and open password prompt on login. You can't pair/connect bluetooth from the quick menu. Focus-stealing is still a thing. You can disable it, but then it breaks getting focus by clicking notifications, which is arguably worse. The file explorer has no title bar, and the navigation behavior is completely foreign. Can't snap quarters of screen. When switching windows (like with alt + tab) windows are grouped by app.

discord screensharing is cuz discord uses an old electron which doesn't support pipewire, the new way to screenshare you can connect to bluetooth devices in the quick menu as of gnome 44, i don't think you can pair tho you can have alt+tab not group windows by app by following the instructions here: https://blogs.gnome.org/fmuellner/2018/10/11/the-future-of-alternatetab-and-why-you-need-not-worry/

Oh that one seems to be fixed somehow

Oh right I did follow those instructions.

Ye nvm about pairing- brain grabs easy words

What about focus stealing?

idk about the rest

Thank you

The smarter the chat I'm in. The less active people in it. And in chats of not high intelligence, activity is always very good. That is, the chance that smart people will want to communicate because of their personal qualities is so small that I usually have no one to communicate with at all — Do you know about this...? — Yes, I know. — A... Okay.... — Marvel or DC? — Yes, it doesn't matter to me. —avif suckers agree? — I believe that everyone does something based on their tasks and goals, I don't think it's possible to offend entire development teams like this. ###Just sit still and work, there's nothing to talk about. And as a result, hiding behind a thick layer of corporate ethics, it turns out that... the only "smart" topic becomes - "what do you think about what doesn't work, although it should work"

<:Thonk:805904896879493180>

And as a result, here are example of a topic that anyone in this chat would like to say - why discord, as a multimillion-dollar corporation and hundreds of millions of users, is engaged in such bullshit as transcoding files into jpeg and webp, instead of ADD new technologies. Do they have specialists?

Well, it is not necessarily about how smart people are but the niche I guess.

But yeah, it looks like you can lose the majority of people pretty quickly

If you got a topic that is overlapping several niches, you can quickly go down from maybe two digit million amounts of people that actually do something regarding each of that topic (with maybe a one digit amount of people participating in communities) — and that is already cutting away the vast majority of humanity that is for some reason not interested in doing anything regarding those bigger topics — to a few hundred people participating. No matter the actual complexity.

My theory is that there is 2 kind of people here : those who use this Discord to synchronize their work (ie. the core developper team and the developpers of other implementations/tools) and communicate news/events/thoughts around the jpegxl and those (like me) who are here to enjoy the discussions, read the news and share the thoughts but have no amazing information to add to the flow... hence, most of the time I read this discord quietly...

I would enjoy to discuss encoding and mangas with you dzgas but I have a newborn at home and my bandwidth to experiment new stuff dropped to zero 🥲

I suppose there are 2 factors. People with interests usually have more things to do. And usually the more intelligent people are, the more abstract and context-depended conversation becomes. Like a single word or short phrase is enough to understand the idea.

https://github.com/JHRobotics/softgpu nice

(hardware-accelerated OpenGL/DirectX/3dfx in Windows 98 VMs)

And yet no one caught on to this topic

To give an unorthodox answer: Incompetence at Discord.

Or more specifically, lack of competence in the domain of image processing.

Those extremely big companies tend to have a lot of people that are incompetent, lazy or busy enough to ignore sane stuff. And those work on the big things.

Well, they could pay people for helping them in that domain.

This has a longer backstory btw. I use Discord since 2016 so I have seen a lot of very bad things

First they need to recognise that there is a problem, which requires a certain degree of domain knowledge.

Yeah, or at least people telling them there is a problem. The latter show up over time tho.

Wouldn't it be also more business related? Count how much money change cost, count how much money it can save/bring?

Well. Maybe true for some companies?

But I know a lot of companies just go the easy looking route when it comes to some decisions

Yes. It makes sense.

Even tho some people inside should know that something is more sane and would save money over time

Like you know, I have only small projects, but sometimes it really takes a lot of efforts to change something. Now multiply that by a thousand to get Discord scale and people just don't wanna deal with it.

I spent several hours on this on my own work. But maybe it will be useful to someone. GDX Texture Packer -- is the best software that makes the smartest of all that I have found packing hundreds and thousands of images of different sizes into one image in the most efficient way that I could find on the Internet at all. https://github.com/crashinvaders/gdx-texture-packer-gui

what do you use it for ?

the world is small, a former author of this tools seems to be a friend of mine from 15 years ago when we both did inline skates 😄

well, that's finely packed ! can jpegxl make use of patches with this kind of image ?

jxl the format certainly can — question is if libjxl the encoder can 🙂

🤨 I would like to understand why I need it at all

https://youtu.be/_4dxA8FUW1w

been a bit obsessed with this lately

Damn, the possibilities. Emotes and text with same font...

~~Xerox gate 2~~

in fact I was looking if in some real-life cases like this one, the font rendering could allow an algorithm to find patches matching but is seems that once again the font ~~kerning ~~ hinting screw up (almost) everything

I would have suspected the (lack of) hinting

yep, wrong choice of word, I found this interesting piece of text : https://pandasauce.org/post/linux-fonts/

ah, I remember the PDFs (some of them) with weird rendering on Linux

embedding the fonts into the PDF seems to solve the issue

whenever I would come across a problematic pdf, I would run `ps2pdf -dPDFX input.pdf output.pdf` and open `output.pdf` instead

> Funnily enough, in my adventures I learned that some people prefer the fuzzy and blurry rendering of OS X defaults on Windows. They found a way to replace the Windows rendering engine with FreeType for this purpose and the tool is called MacType. Since FreeType is highly configurable, it allows you to disable hinting and mimic OS X approach to rendering fonts. You will see below why this actually makes for a worse user experience without even talking about stability issues that this is bound to introduce. I use MacType and I have not seen below why it makes for a worse user experience

perhaps you got used to it 😄

not just that – there must be a reason I started using it in the first place, no? 😉

it was not force-installed on my machine

but then, at 4K/32" (138ppi), I guess my screen is “mildly high-dpi”

to my taste, and despite my aversion of patents, I found the progress brought by clearType was really a game changer at the time

indeed, the issue is dissolved by high resolutions

Qt6 has noticeably improved high-dpi font rendering on Windows compared to Qt5 (at least with weird scaling factors like my 150%):

(in this case, it seems that disabled groupboxes are no longer rendered gray so perhaps I should have checked the checkbox so that the text would be black in both cases)

the difference is very noticeable !

gtk2, are you okay?

https://twitter.com/infectio01/status/1606355717371281410

Lol why is it like that?

I loled at "roar" ngl

https://twitter.com/Google/status/1653866291692728320

why google, why .zip

What is the point of all these new silly TLDs? It has been going on for a while now.

I don't mind more TLDs I just find it weird that they add .zip

that's like, asking for phishing to happen

The branded TLDs are the worst for me, cluttering up the namespace with parked top-level domains

Arguably they don't clutter the name space in the same way, as they are more creating new namespaces.

But when they are really specific it does make you wonder about their longevity. Cool URLs don't change. (https://www.w3.org/Provider/Style/URI)

.mov too

Something like `https://apple.com-firmware.zip`

`https://zip.zip/zip.zip`

I think zip has now lost all meaning to me

.com used to be as common as .exe in DOS

Note also that there *is* a `.sh` TLD. 😅

Yeah but as the country TLD of Saint Helena, created under ISOs country code system

I think there is also less trust for something ending in `.sh` than for something ending in `.zip` or `.mov`, which is important for the use of a domain in phishing

the technical term for this is https://en.wikipedia.org/wiki/Semantic_satiation

This will be my first time trying to use extended partitions.

as in, the workaround for the MBR 4-partition limitation?

I just use gpt now

Yep

Why are you using mbr

who doesn't? 😛

MBR is useful for firmware flashing (usually when combined with FAT32), but I don't really see much of a use case for it outside of that these days.

I used GPT even before switching to UEFI

(syslinux supports BIOS/GPT)

I use Grub instead of syslinux mostly cause it's what I've always used

if you have no need for partitions, e.g. portable USB drives, then MBR + Fat32 is universally supported

I don't really see the need for GPT for pluggable non-bootable drives

well, >4tb, no?

(granted most pluggable non-bootable drives are not >4tb)

there was a time of my life in which I just used efi-stub as a "bootloader"

my favourite nowadays is rEFInd

ye I don't see 4TB removable usb drives

the netbook I bought in 2012 turned out to have an UEFI firmware disguising itself as a BIOS

it normally wouldn’t expose the EFI functionality

I have a 32tb "removable usb drive" 😛

unless there was an EFI bootable drive plugged in, in which case the F10 (?) boot menu only would offer to boot it

I never found a way to default to an EFI boot

but I did make it so that my netbook’s hard drive could be booted indifferently in either BIOS or UEFI 😄

I don’t remember if I ever tried hibernating in UEFI and resuming in BIOS or vice versa

I think I might have, but then I don’t remember the outcome

Idk if I'm actually using that. I saw "4 partition limit" and said yes.

MBR is what would you'd be using if you had a 4 partition limit

GPT has no partition limit

(well, effectively none)

255 on linux and 128 on windows

idk about mac os.

effectively none <:kek:857018203640561677>

16 or 32. https://apple.stackexchange.com/questions/80794/why-should-i-observe-a-limit-of-sixteen-partitions-per-disk-with-os-x

well effectively none on other systems ig

that link is from 10 years ago

things might have changed in the meantime

https://www.instagram.com/reel/Crg4XiUswX6/?igshid=NTc4MTIwNjQ2YQ==

awww

just realised that macOS’s GUI allows naming files with (apparent) `/`

they are stored as `:` on the filesystem

and so it’s `:` that the GUI disallows

???

(what led to this discovery is that I tried to save a file as `/tmp/a.jpg` and ended up with `~/Downloads/:tmp:a.jpg`)

What happens if you try to name it with `:`? 😄

it tells me that that name can’t be used and to try a name with fewer characters or no punctuation marks

> Essayez d’utiliser un nom comportant moins de caractères ou ne présentant aucun signe de ponctuation.

Huh. I guess for the end‐user, wanting to name a file with a slash is more common.

Though I think it’d make more sense to have chosen an even less common punctuation mark as a surrogate like that, perhaps `~`.

that is so cursed

And about 26 (or maybe a few less even) during the Windows installation (not a joke, it will refuse to install with a barely helpful message if a drive with more partitions is connected...)...

_Me with a few dozen of partitions on one HDD:_ Yes, ha ha

https://cdn.discordapp.com/emojis/933991733237276692.webp?size=48&name=CB_shibaheh&quality=lossless

i thought it was a weirder character

my brother named stuff with mm/dd/yyyy on a flash drive and i recall having to use tab completion to enter the filenames on linux

`:` isn't permitted in DOS which makes it weird on FAT32 filesystems

so it would be something else

this is awful https://blog.google/technology/safety-security/updating-our-inactive-account-policies/

Seems to be happing at a lot of sites all at once

I will be trying to do a lot of backups while I still can

I can see where they are coming from, but I don't want to lose the absolute treasure trove of conference talks uploaded by random people 10 years ago, etc.

If I were they I'd "fossilise" inactive accounts, making them read only, or even just remove the ability to log in.

Though I do wonder if that could cause legal issues around the right to deletion; you wouldn't be able to delete an account that had been fossilised.

Simple rule. If you want to keep something important, make at least 1 backup.

false alarm I guess? https://twitter.com/YouTubeLiaison/status/1658829489466384387

> When you purchase one of my courses directly from my site, you have 24/7 access to it for as long as this site exist. I’m sure this is meant to be reassuring but I’m not fully comfortable with the statement

"You have unlimited access, until you don't"

it’s a course about another kind of Modular

(modular synthesis)

the course uses this software synth: https://www.arturia.com/products/analog-classics/modular-v

a software recreation of the classic Moog

Nya!~ :3

UwU

<:UwU:911376902411919421>

Top is real, bottom is malicious

Ah, the slashes are wrong! 😄 Took me a while to notice.

I’d expect this would be resolved by virtue of showing it in punycode form, like browsers would likely do anyway.

But isn’t it the other way around?

Ah, it’s not a domain name, it’s a username!

And indeed, the top one seems to be the malicious one. (I.e. you got it the wrong way around.)

Thankfully, browsers (and I think most tools) will elide the credentials from URLs altogether.

Where the top one is malicious because the domain name is `v1271.zip`, and the username of the URL is `github.com/.../tags/` (with funny Unicode slashes).

Just goes to show how confusing it is

looks like japanese slashes

Top is real bottom? 😳

／

They are all slanted lines :／

https://twitter.com/vzverovich/status/1659687358474891265

Less efficient than XML <:monkaMega:809252622900789269>

https://github.com/64/cmake-raytracer

What even is cmake

A system used to build C/C++ projects, it comes with it's own script language to cutomize your builds

Cmake is a huge spaghetti mess pooped on top of the pipe nightmare which `make` is...

make isn’t CMake’s only backend

I mainly use the ninja one nowadays

https://github.com/libjxl/libjxl/blob/91760e31a40fb811867fbe9bda6ddf73c962389e/ci.sh#L343

just did a `for d in */; mv $d (echo $d | perl -pe "s/...//"); end` (fish syntax), on windows, and ended up with directory names ending in `\r` 😩

```sh for d in */; do mv -v "$d" "$(printf %s "$d" | sed 's/...//')"; done ```

trick is to use `printf` instead of `echo`

or to use `"$(foo)"`

which helpfully strips all trailing newlines from output

including CRs on windows

I’m not sure fish has an equivalent to "$(...)"

oh this isn't shell?

I mean, it’s supposed to be (...), but it seems the windows implementation doesn’t strip \r then

either way `printf %s "$d"` won't attach a newline at the end

`echo "$d"` is unsafe anyway

since echo accepts switches

not that you should name a directory something stupid like `--help` that's just asking for bad stuff

true, but I usually don’t name my files with starting `-` anyway 😁

(these were my files)

but `printf %s "$d"` is resilient against that

I ended up fixing the problem with a quick perl loop

but ye, in the future printf is the way to avoid a newline at the end

e.g. if you want to md5 something from the CLI

you can do `printf %s "$variable" | md5 -`

which will md5 the variable without the newline

My jaw dropped when I saw this woman wearing a pattern similar to a pattern that I discovered

The right image can be generated with C code like ```c image[y][x] = (x * x ^ y * y) & q ? BLACK : WHITE; ``` for a range of `x`s and `y`s and some `q` that is a power of 2

That doesn't _look_ like it has a short jxl art representation, but maybe there is...

In any case it's probably a very effective pattern to create Moiré

Ninja'ed by gucci

https://twitter.com/panley01/status/1661463308971417601

I didn't get the reference 😅

hehehe

A320 doesn't have a pit 1

that's all lol

but what is pit ? like brad-pit ? cock-pit ? pit-ot ?

it's a compartment in the underbelly of the plane

oki

didn't know 😄

ye that's why i love that meme lol

https://i.redd.it/vm53mf20vx2b1.jpg

I hate this

if not ii w then w

looks like someone's been defacing our poor city's buildings we'll just have to teach them a lesson

<:CatBlobPolice:805388337862279198>

Bruh

Do you use arch btw?

does people really _use_ arch ? I heard it was only meant to recompile package, reinstall it again and again and spend all your week-ends to configure it...

Bruh

Arch is not gentoo

😄

yeah, no, you’re thinking of NixOS

I have never had to reinstall Arch on any of the machines I’ve installed it on (4 + 1 VM)

(although it came close, once)

I personally use arch although the meme about "I use arch btw" is a bit excessive

NixOS and Gentoo are more beardy

and yea Arch is far more stable than it was in, say, 2012

recursive `sudo chmod` on home directory + the `~/.wine/dos_devices/z:` symlink to `/` do not combine well

haha, that one is user error tho

since there is a `--preserve-root` option

`--preserve-root fail to operate recursively on '/'`

I don’t think it actually operated recursively on /, just on / itself (which made the diagnosis harder)

`ls -l /` didn’t show anything suspicious

ah I see

_but_ `ls -ld /` (or `stat /`) did

but in all seriousness, arch linux has come a long way when it comes to stability it's pretty usable as a daily drive

less than LFS 😄

I used it 2010-2020 (and still in a VM)

what do you use now?

I migrated to systemd before it was even mandatory

Windows and macOS 😂 😅

for development I find POSIX is easier

than windows

not sure how compliant macOS is with POSIX, tbh

it's freeBSD under the hood... but

nah, it's not based on FreeBSD

I find that msys2 and the new windows terminal alleviate much of the pain there used to be with windows

I prefer WSL over msys2

and yea the new windows terminal that you can actually make bigger <:KEKW:643601031040729099>

and with tabs, and better support for non-ASCII characters

and clickable links

etc.

and curl

:D

you used to have to use powershell to download a file from the internet with the windows CLI

no, I use linux for 1 year and then switched back to windows, to the repack windows, to the windows that were turned inside out and deleted everything except Windows

I just don't like arch users.

why not?

How do you tell someone uses Arch?

you ask them what operating system they use and they tell you

Nah, they'll just tell you

yes I'm aware of the meme but it's really not that true

it's really just a meme

it's pretty silly to dislike everyone who uses a particular distribution cause you believe in some meme

You're right, it is.

https://github.com/crablang/crab/issues/59 🔪

bruhhh This is epic

when steamdeck

what's steamdeck

There's your awnser

LOL

What's the difference between Crablang and Dust?

atm there's no difference between Crab and Rust, it's essentially a fork with no functional changes, just renamed

it's a protest to the trademark controversy

No no- Dust, not Rust

uh, I'm not familiar with Dust so idk

what AI software is this <:KEKW:643601031040729099>

looks like SD

yea but the UI looks different

o

I was wondering what the frontend was