I was at the talk, and it's strange this what people take from it. You should watch the whole thing and see what he built over years.
I was a bit disappointed that most of the questions ignored his talk about a very cool jukebox he built and focused on OS drama.
He built a jukebox with all hit songs he could find in it 1900-2000 and for prerecorded music, got a player piano and sheet music and midi and integrated the whole thing. Touch screens, voice activation and so on. Hardware and software and data hoarding project.
He said he has massive cabinets of CDs, all the music he ripped and tested audio encoders with his own ears.
Ken is 80, and still building cool side projects and scratching his own itch! That's the story.
Be like Ken by building something cool, not by using whatever OS.
I was there too and, at first, was wondering if the music thing was just a warm up to the main talk. But soon enough I settled into the flow of it.
His slides were incredibly minimal throughout and then, at the end, he played a video of maple leaf rag pouring forth from his player piano midi setup and it was like choirs of angels singing.
How many boring ass talks have I sat through that I’ll never remember — but I suspect I’ll remember his talk for a long long time.
Edit: I've put a placeholder up there for the time being until we get a better suggestion. Submitted title was "Unix legend Ken Thompson announces he's switching From macOS To Raspbian Linux". I agree with the parent that this is trivializing (also cherrypicking and editorializing) and we should focus on the substance of the talk.
I listened to the question and answer at the timestamped link, and I wonder if he was giving a completely bogus answer to see if anyone was paying attention. Consider:
> I have for most of my life, because I was sort of born into it, run Apple.
Assuming that's actually still Ken Thompson talking, that makes absolutely no sense. He's several years older than Woz, never mind Apple itself. He was already well into developing Unix on DEC minicomputers when the Apple I came out in 1976. Then, all through the 80s and 90s, I'm sure he used whatever non-Apple computers they used at Bell Labs. I think I read somewhere that by the 90s they were using x86 PCs. Anyway, you get the idea. So I wonder if he was totally messing with us in that answer.
Thompson's talk is very interesting, but it's also interesting that Thompson has decided to use a UNIX clone for his daily compute over macOS. The original submission included the timestamp to concisely link to the comment. It might be that autocanonicalizing YouTube submissions to remove the timestamp is the actual issue, perhaps that could be reconsidered.
The one is a talk going deeply into a cool technical project and the other is a cherry-picked bit of celebrity gossip—admittedly a fun fact, but not interesting (at least not intellectually interesting) in the same degree.
Although we didn't write code to do it on purpose, I think removing timestamps from videos (and similar things, like removing HTML anchors) generally does more good than harm, because people tend to use those to cherry-pick some detail they think is important rather than letting the reader make up their own about the whole submission. Generally we try to discourage that (see e.g. https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que..., which is about titles but makes more or less the same point).
Well that's certainly your call but it's not gossip when it's first party at the very least. It's a statement by a luminary in the industry. As for timestamps, I can understand the intent with stripping them, but it might be an outsized measure when there are definitely other cases where including timestamps can help with context. Perhaps most usefully, allowing timestamps ensures the submitted title (which we as readers are voting on) matches the actual content you get when you visit the URL, so that the topic of comments can remain aligned with the submission itself.
Ideally, I think this talk should have been two submissions, one for the OS commentary (which HN mods can choose to moderate or not) and one for the actual talk.
dang, it's far too late to change the submission title when there are already hundreds of comments focused on the old title. The commentary makes no sense now.
If anything, there should be a separate submission with the new title.
Normally I would agree but not in this case. There's plenty of context in the thread, readers are smart enough to figure it out, and adamgordonbell is right: the talk, and the speaker, deserve better.
You're right that titles dominate discussion though. That's one of the most reliable phenomena on HN, for better or worse.
i submitted it ~4 days ago with the title "Ken Thompson's 75 year project", but the title was evidently too boring to gain any traction :) (like some others have commented, the raspbian os "revelation" was the least interesting part for me too)
it was cool to hear that he's still working on chess endgames, an interest which goes back many decades too [1].
I have to say I was disappointed by the question at 59:23. They seemed to expect a retrospective on Ken’s career or some grand philosophical statement on software or open source. To be honest, I was pretty surprised by the direction of the talk myself, but I ultimately enjoyed it.
You see, Ken decided to talk about his 75 year project: his music collection. He talked about audio formats, collecting music from different groups, sourcing metadata, building hardware to play music and more. He was deeply interested in the topic and honestly probably a bit obsessive for multiple decades. This was very humanizing. And to be completely honest he reminded me a lot of my girlfriend’s father who we think is undiagnosed autistic.
Ultimately, I think the reason why Ken was so prolific over such a long time is his ability to be deeply interested in problems. He was not too fussy about tools. He didn’t push Go or Linux or UNIX. He wasn’t self aggrandizing. He just wanted to tell people about his project that he’s been working on. Honestly, I thought it was a great lesson that might have gone over a lot of people’s heads.
I was there in person and I debated with whether or not I should shout down the guy who basically insulted him for giving what I thought was a great talk. Obviously, I didn’t (because nothing good would come from making a spectacle like that), but it bothered me that he was insulted for presenting what I thought was a really beautiful and humanizing talk.
I met Ken the day before and he told me he was nervous about his talk. I hope he knows that most of us in the audience really loved what he shared.
That part kind of broke my heart. In retrospect, I should have gone with my gut and shouted “we love you, Ken” after he did that. Because you’re right, he doesn’t owe us anything. And he certainly doesn’t owe us trying to uphold some ideal we have of him that we’ve invented on our own.
It's a technically oriented, open source, Linux conference and he's talking about basically his home stereo.
Well ok, I guess he can do whatever he wants. There's expectations from the context though.
That speech was the first time I understood how people can become too famous for their own good.
I saw him the day before (and I've got the photos to prove it) and he made a bunch of factual mistakes. Nobody cared enough to push back. I only did a little when he claimed raspberry pi prices have been stable and plentiful throughout the pandemic. But even I quickly gave up and I'm usually an asshole about things being right. (2019: 1gb pi4 was $35.00 and now it's $132.95 for the record. Buying 1000 cheap computers to sit on them 4 years and then just resell would have made you about $100,000)
He can basically say whatever and do whatever and people just politely murmur in approval because he's computer royalty. There's no real feedback loop or anything to keep him in check.
It'd be like if Beyonce held a concert and then instead of singing, talked about composting and gardening. I mean sure, whatever.
The man is 80 years old. High status doesn’t give people a pass, respecting what people have contributed and not having the need to “well, actually” every little thing does.
I met him Saturday too and I was struck by how lovely and down to earth he seemed.
Who cares if he isn’t intimately familiar with RPi4 prices and availability? The cyclical and location-based nature in how these things work makes it impossible for anyone not absurdly obsessed to keep track. (I bought 8GB units in 2020, 2021, and 2022 at list price. If I wasn’t trying to get some CM4 units, I wouldn’t even be aware there is an ongoing shortage and I consider myself to be incredibly plugged in)
I found his talk about his home stereo to be far more interesting than a talk about the inner-workings of UNIX that’s he’s undoubtedly given a thousand times before. To each their own.
> Who cares if he isn’t intimately familiar with RPi4 prices and availability?
He says he has about 100 of them and made a number of specific claims on it - that he's some huge supporter of it. I chose that example because it was conclusively wrong.
Regardless, that's not the point.
The point is everyone will rise to his defense based on his status regardless of his content.
Not like that has been clearly demonstrated in my replies or anything ...
He didn't know pac's relation to he-aac, or digital FM was renamed hd radio (which uses mdct, based on he-aac, based on pac). I was volunteering for the debian booth and thought Debian and raspberry pi os had no relation.
None of these things matter because you're just being an asshole.
I think being obsessive that someone didn’t know digital fm was renamed he radio or that Raspberry Pi OS was based on Debian (maybe if the name was still Raspbian it would be more obvious) makes you more of an asshole than the guy you’re responding to.
I had to Google this. Seems it is some US-only thing I've never heard of.
The rest of the world calls digital radio "DAB" and occaisionally "DAB+". I've bought 2 DAB radios so far this year, since moving back to a country where I speak the native language.
I recall reading an interview many years ago where he mentioned downloading large quantities of music to build his personal collection and that his employer, AT&T, did not object.
> Community sourced headlines (toggleable visibility natch) would be a great addition. Tie the voting to karma on them.
I remember when I first discovered the StackOverflow website. Every question had tags on it and some number next to each tag. Me at the time thought that this number meant how many times people had added each of those tags to the specific question I was looking at. It took like a year or something before I realised that this is not what those numbers meant.
“Unix Legend Ken Thompson, still out of step, taps his foot to the sterile sound of CDs just as they are now again being outsold by warm, pressed, analog vinyl”
He has already been using raspberry PIs for years, and if you watched the talk you will find this has nothing to do with iTunes, it predates it by decades.
today it is. tomorrow it might not be at all. or it might be $100/mo. or they might decide they need interstitial ads. or only if you use their recommendation engine with preselected 'channels'
the computational requirements to store and play music are so minimal now, I'd rather just take care of it myself
Apple music is lossless from the moment lossless analog signals get imperfectly sampled and chopped into bitly approximations, yes. There must be an alias for the name of that process...
(i love DSP, calm down, but i also like lossless English)
what about designing the DSP/memory scsi disk that managed to do realtime ripping to PAC, then changing all that to mp3 using LAME, and the most important, creating usable metadata and finding all these CDs, and even finding ways to find a top-N song collection from dates that didn't have any relevant publications. I think you're underselling his achievement a bit ;).
I found the talk to be an epic geek journey.
Thompson: It's kind of a personal/research hobby/project. Let me explain it from an external point of view. Basically, I'm just collecting music. I'm getting lists from various sources—top 10s, top 50s—and I try to collect the music.
Right now, my list has around 35,000 songs, of which I've collected around 20,000. I compress the songs with a Bell Labs-invented algorithm called PAC [Perceptual Audio Coding] and store them on a jukebox storage system. I started this before MP3 was heard of on the network. PAC is vastly superior to MP3.
My collection is not generally available because of the legal aspects. I went to legal and told them I was collecting a lot of music, but I don't think they realized what I meant by "a lot." Anyway, they said that in the case of research there's something similar to fair use and that they'd back me, but wouldn't go to jail for me. So I can't release it generally. But it's pretty impressive. It's split-screen like a Web browser; you can walk down lists, years, or weeks.
Computer: It's a personal hobby.
Thompson: It's hard to differentiate since, if you haven't noticed, almost everything I've done is personal interest. Almost everything I've done has been supported and I'm allowed to do it, but it's always been on the edge of what's acceptable for computer science at the time. Even Unix was right on the edge of what was acceptable at Bell Labs at the time. That's almost been my history.
I feel privileged to watch this. I live in north Africa and I feel like taking a flight to California to go visit him, inquire about his health, tell him about my children progress at school, show him how much I love him and how much I'm grateful.
completely agree with you. ken and the not so big group of people who has positively shaped all our technology really need more appreciation from us users, although i'm sure he would be weirded out by it like any proper geek. That been said: thanks ken!
It's also worthy to note that shortly after that part of the video, he notes (from another question) that he has over 50 Raspberry Pis (including 12 stacks of 4xPi4s). So his choice of Raspberry Pi Linux is likely the result of that.
The announcement is in the Q&A after the talk - but the talk itself is definitely is worth watching. It starts at 10:56 (link below), and covers his "75-year project". It's kind of an amazing story that his life has spanned so many different eras of technology.
I was using a combination of windows/linux for a while until my archlinux laptop shit the bed after an update and I decided to say, fuck it, I'm finally buying a macbook because at least then I can still do unix shit without having to worry about everything working the next day.
I'm not happy about "Apple Silicon", it does feel restrictive and often times the only way to get around it is to use licensed VMs, which feels like a bit of a rip off. At the same time, my laptop runs phenomenally well, does everything I need it to do, and it never dies or gets overheated under normal use. I can't really complain.
Archlinux is a rolling release distro. Which mean you are supposed to:
- update them very frequently
- check the news before any update which mention any manual intervention needed
This is definitely a distro for people who want to get involved in the sysadmin part of it. A distro like Fedora will have a new release every 6 months and each release and ~390 days of support. You aren't expect to do manual intervention, just let it update itself upon reboot/poweroff from times to times and do a major upgrade, every 6 months to 1 year. Debian has 2y releases cycles and +- 3y of support, in the ubuntu world it is like Fedora or LTS (5y of support). If you want the longest extended lifecycle, RHEL and its clones have pretty much 10years of support without any fuss. Add to that the immutable distros like Fedora Silverblue, OpenSUSE MicroOS and some others where it is virtually impossible to make it shit the bed, even while being stupid.
So in the linux world you can definitely choose your poison, from the less eventful one to the one needing more attention. It looks like you didn't choose wisely. If I had to setup a distro for the least knowledgeable people, I would set it up with an RHEL or Almalinux and install a more modern browser through Flatpak. As long as the hardware is supported from day 1 they would expect a desktop that do not change at all for 10years.
I would love to know more about what you mean by 'archlinux shit the bed'. If you dont remember the exact error code, any text from the error screen, or even vague description is fine.
I believe that spending 15mins on the arch forums or IRC would probably result in somebody helping you out with the right pacman incantation. YMMV, my personal experience.
On a tangent, I would love to hear more about any Debian stable users out there and their experience with the conservative approach to updating.
I am particularly fond of this Debian wiki article on DontBreakDebian: https://wiki.debian.org/DontBreakDebian
You might like your macbook's hardware more than you think. It's a pretty open platform with a standardized instruction set and on open bootloader than can run Linux with pretty good hardware support for such a new platform.
So you went from archlinux, a system explicitly telling you you have to get involved in your own computer to make sure it works, to a macbook where everything is taken care of for you. What did you expect with archlinux exactly ?
That is a hacker's hacker. Hacked a 50s jukebox that combines LCD display with manual switches and supports voice input to play the chosen song on a player's piano - from a catalog that spans a century.
Also loved the video of his wife enjoying the setup - straightforward and effective.
What an incredible guy. Him, Dennis, the rest of Bell Labs, and all the other less known influential computing pioneers are such treasures. We're a fortunate field to have the kind of people that we do.
To those who are complaining about the lack of power/issues with Raspberry pi, don't forget, this is not one of your run-of-the-mill tinkerers, this is Ken Thompson.
If I'm not mistaken, Apple was the one that blacklisted Nvidia hardware, not the other way around. You can still download bog-standard UNIX drivers for Nvidia hardware, Apple just won't let you install them.
It was understandable that Apple stopped shipping Nvidia hardware. Blacklisting Nvidia hardware and preventing the installation of Nvidia drivers is another thing, though, and goes too far in speaking on the user's behalf. If that's somehow more of a "narrative" than Apple's perspective on the matter, I don't know what to tell you.
Can you event change the GPU for an nvidia on any Apple device? If you are talking from a hackintosh perspective, I understand they are perfectly allowed to not caring about that.
I just wish there were hardware to support a full switch to Raspbian. Ken has a lot of RPi 4. Those are usable as an interactive desktop but it's not a great experience, the hardware is just barely capable of being a responsive desktop OS.
I really like what Google has been doing with ChromeOS and Chromebooks. I wish there were a program like Chromebooks for a Linux desktop. Arguably that is ChromeOS itself, but the Linux environment you use is a VM.
Have lived with MacOS on a “late 2014” Mac Mini until it became so slow as to be virtually unusable (amongst other things). Now happily run Linux on it.
FWIW I still use a 2011 Mini and my fiancee's laptop is a 2009 MBP.
The one thing you need to know is that, just as with Windows, occasionally, you must reformat your machine and reinstall the OS clean.
Do not migrate your files & settings. Backup your files, yourself, manually, by hand. Clean install the OS, reinstall the apps you need, yourself, manually. Then copy back your data.
If you use an automatic backup tool, it also brings all your cruft back. Don't do that. Do it by hand and only keep the stuff you need, not what some automatic wizard app thinks you need, because they don't think.
It is, in my experience, much less of an issue with Linux. Helps there occasionally, too, though.
Old Macs are still perfectly usable if you do your housekeeping. They have not mystically slowed down over the years or anything.
Exactly. Everyone is missing the big picture, focusing on Ken's jukebox, when its pretty shocking that Ken's giving up on a BSD-inspired (UNIX v7-descended) OS. Frankly, I was ready to throw my employer's mac against the wall when it refused to allow me to install gdb. The move towards gatekeeper and putting everything inside its own private walled garden will definitely ruin open source tools that can access any file on the system, making apple superior to all developers. I am pretty sure that some of these gestapo changes have been rolled back.
I don't really want to get too deep into it because this same discussion comes up every time there's an article about FreeBSD :)
But my personal reasons are:
- Less commercial influence on the OS and "distro" development process (think of things like Canonical pushing snap and other not-invented-here behaviour)
- A stable (release-based) base OS with rolling packages. The perfect combination which for some reason is not as common in the Linux world. There it's usually all-rolling or all-release.
- The ports collection - Recompile any repository package with any parameters you like
- Excellent documentation because the world is not as fragmented as on GNU/Linux
- Not as much drive to constantly change things as on GNU/Linux (which is partly driven by point #1 of course)
I did exactly the same and my reason is : Linux is chaos, FreeBSD is order. FreeBSD is so very well engineered, everything feels just right and logical. It's not the case with the fragmented world of Linux distributions. Unfortunately FreeBSD doesn't have the laptop driver coverage Linux has so I'm using Void Linux on my laptops because it is the most BSDish Linux.
Yeah good point. I'm not really a 'laptop guy'. I only use desktops (mostly NUCs). Which is basically a laptop without the battery, screen and keyboard anyway.
But I've heard WiFi drivers in particular are not so good - never really looked into it because I wire all my stuff up anyway.
I use a thinkpad laptop for work and a cheap $150 Chuwi laptop for the makerspace but that's all.
I’ve given strong thought to switching away from macOS. I too have been a Max user all my life, a Macintosh Plus being the first computer in our house. I would get fed up at Apple’s hardware choices or its limitations on users.
I have had a dozen Linux computers with various systems on them. I don’t know if it’s because they were Dell machines, or if it’s an Ubuntu thing, but I have had almost every single one turn into a brick after a Canonical-issued update.
The kind of brick where you have to boot into the boot loader and into single user mode (?) and start issuing arcane commands to try to recover your system with some old kernel.
The thing that keeps me on my Mac is that I can mess around with Unix computing all day, and then go back to being with done when I want to get back to using my computer. I don’t feel confident like that with Linux.
My experience is that what holds Linux back is NVidia. Their proprietary drivers work great when you first install them, but inevitably break on update bringing you to a text mode emergency command line. When I made the switch to Linux in 2018 I made the conscious decision to avoid Nvidia hardware and it has worked out really well.
NVidia have moved most of the driver stack on to device local firmware instead. The parts of their driver that interface with the GPU are now largely open source as a result.
Of course this only applies to more recent GPUs so doesn’t invalidate your comment
Only if you think that firmware doesn't "interface with the GPU", 2 biggest parts of their driver are that and profiles for games/apps (which are accessible by design though), and they've open-sourced the loader of firmware recently.
I'm sorry to hear about your experience! I feel like noting that my primary driver has been Kubuntu on an (nvidia) laptop for a few years now, and it's been the most pleasant experience. Sure, you get rough edges every now and again, but I was honestly getting rougher edges on the Mac (it's been a long time since they 'just worked', alas).
Certainly no involuntary grub prompts to date, thankfully! Happy as a clam with my Linux laptop as a daily driver, including for gaming (!) and work.
You might like Guix System or NixOS for their concept of "generations" and easy rollbacks. It's pretty hard to break your system majorly. At worst you roll back and then figure out what broke (when you feel like it), and then you can even do partial upgrades and skip certain packages if they caused issues after an upgrade.
Here I am waiting for apple to finally introduce side loading (and with that hopefully easier ways to jailbreak) in iOS so that I can switch to a better smartphone experience.
There is zero reason to assume that sideloading will make jailbreaks easier. Anyone can already sideload today using a tool like AltStore [0]. It uses the same mechanism that app developers use to test their apps on actual hardware. The only difference to “real” sideloading is that Apple limits it to a maximum number of ten apps and that the apps need to be refreshed after seven days.
Apple claimed “Unix” because Mach shipped with BSD tools for testing and research purposes.
Now even Windows has a Linux compatibility layer now too and everyone but RTOS all run containers and VMS.
Funny story: even Apple switched to Linux in the Data Center. They doubled down on Appliances (which happen to run an OS).
Are people pressed that an appliance won’t fit hacker workstation/embedded needs!?
Well good for him. But we have to let go of this notion that for GNU/Linux to win, Apple has to lose. I'm paraphrasing but can't shake that feeling when I hear those claps in the video. When are we ever going to grow up?
I'd summarize it more as him saying he's choosing to opt out of Apple's direction because he doesn't like it. Was there something beyond that...calling for Apple's demise?
I was in the audience and I certainly didn’t clap. It wasn’t the whole crowd (tho it was most of it, as one would expect at a major Linux community conference), but I rolled my eyes by the crowd reaction too — tho I don’t blame/judge/fault people for their reaction. I’ve always been the Mac user at the Linux conference (and I like Linux very much, but I like it on the server or an RPi) and you know going in it’s like wearing Yankees hat to a bar full of Red Sox fans.
Hehe, yeah. I saw Stallman in Gothenburg 2007 and even bought his book Free Software, Free Society. It was memorable, getting flashbacks now from him eating toe jam/callous skin from his foot.
At least for the first request, Pop OS is a fantastic replacement for the Mac. It the system 76 variant and is based on Ubuntu. I switched last year and have been blown away at how far desktop Linux has come
I use it full time for work and home. Chrome, VS Code and Steam all work flawlessly. It's also nice to be able to develop software in containers without needing any VM layer.
I have used Linux for most of my adult life: I don't get the sens of control over my own computing with other operating systems. It's not clear though why exactly Thomson is switching.
when asked what language ken used, he seemed to answer "Algol"? I'm surprised there's no further discussion or clarification in the comments. if anybody heard clearer, or has some feedback on what he actually might've been using, I'd appreciate a response.
the other question that I have about the video is, which one is "the alligator talk"?
In case anyone wants to know if he clarified on why Raspbian in particular: he did not.
Raspberry Pi hardware would be the obvious constraint. Though, that only pushes the question one step further: Why a Raspberry Pi hardware? I... don't know. If you went from Apple devices, it seems like a non-sequitur. There are much, much better options, unless on an extremely limited budget.
But, I also don't know his use case. Does he have a gazillion of them for some (I assume interesting) reason?
Raspbian [1] is the name of the original community project to port Debian to armhf (armv6 hard float) which ran on the Raspberry Pi 1.
Raspberry Pi put our their own images based on this, and called it Raspbian until about 2020 [2] when they started calling it "Raspberry Pi OS" after they started producing aarch64 images.
I don't see the phrase "Raspberry Linux" in the linked page.
Google results only seems to show those words as a deliberate name for a few niche things like "RT Raspberry Linux". Meaning, I still don't think "Raspberry Linux" is a thing. HN should probably change the post title here.
He said Raspbian. It's just the post title here that's off.
Edit: Yes, it's something of a nit, but it helps for searching, etc, later. Or if in the future, a different product does have that name. "Raspberry Pi OS" is an option if Raspbian seems obscure.
That tool gets you the option to install the image of your choosing.
8GB can be exhuasted, use a System Load Viewer dock widget and btop to spot/stop/start the process over-consuming your memory. Typically, the web browser hogs up to 1GB then I quit it. The next upgrade is to 16GB or 32GB from 8GB for me. I went from the Mac Mini at 8GB to Rpi 4GB then 8GB.
Ken Thompson was also part of the Bell Labs Plan 9 team before they went to Google and created Go, but I’m sure both don’t use it as their regular OS today.
I watched this whole video, (except the front bit). At the risk of being flamed, the audience came off as being incredibly rude. Perhaps this is par for the course with this crowd; if so, I am glad that I am not one of them.
Q: What's your operating system of choice, today?
A: I have for most of my life, because I was sort of born into it, run Apple. Right now, recently, meaning within the last five years, I’ve become more and more and more depressed and [Laughter] what Apple is doing to something which should allow you to work is just atrocious, but they are taking a lot of space and time to do it so it’s ok. [Laughter] And I have come within the last month or two to say even though I’ve invested a zillion years in Apple, I’m throwing it away and I’m going to Linux -- to Raspbian in particular. [Applause]. Anyway, I'm half transitioned now.
I don't know the exact issues but might be related to the issues why I left Apple : I was developing a very successful little tool for MacOS ( https://github.com/milgra/macmediakeyforwarder ) which listened for keypresses. From 2016 to 2019 it became harder and harder to install it because apple added more and more restrictions to apps like this. By 2019, you had to enable the application explicitly to listen for events at least in three places deep down in the system preferences, click accept in various popups and if you stuck somewhere then nobody could tell why it wasn't working. So I had a very expensive laptop and the OS didn't let me use it freely. So I just switched to freebsd and linux. Hardware quality is far away from Apple's but it is cheap, I don't have fancy productivity apps like photoshop and final cut but with open source tools and with my own desktop applications I created the best looking/most usable desktop experience MacOS will never have. ( https://swayos.github.io/ )
I think permission systems are bound to wind up in all desktop operating systems, eventually. They’re already on Linux for those using Flatpak. Trusting random third party binaries with access to everything is increasingly too much of a gamble, even for more technical users.
That said, I agree that the macOS implementation has issues. It’s tricky though, because if they make it as simple as confirm/deny dialogs, you’ve set users up to quickly succumb to “yes-click syndrome” which is likely why Apple went with the “flip a switch in a preference pane” design for some permissions instead.
“Click yes syndrome” is so prevalent is has an actual name: banner blindness.
Last year you might have heard a court case talking about a nurse who killed a woman by accidentally giving her the wrong medication. She took responsibility but talked at length about how the system in place encourages workers to blindly click yes on alerts about medication because there are many of them. The training they got was basically “just click yes three* times” (* I don’t remember the actual number but three seems right).
One of those warnings could have saved a life had she read it, but she had been clicking yes many times a day every day for a long time and she no longer even saw the banner for what it was.
Insecurity through obscurity is possible even in open source. See log4j, but there are other examples — and infinite proof of concepts of people breaching repositories. Even on the desktop, you want multiple layers of security to limit potential damage.
Do use Linux on the desktop and be happy if it makes you happy, but don’t smugly assume you’re immune to the outside pressures in today’s world that are causing Apple to institute basic UI security measures on macOS. This isn’t a walled garden issue, it’s “make sure the user knows this binary is doing something that allows it to be a keylogger if the developer is so inclined.”
Android app vendors are not writing their own policies. So there is a lot of code between SELinux and "what's actually usable to the user".
Like, stock RedHat does too, it just took a ton of effort (and bugs) to get there.
But then it is complex problem so its no wonder that the tools to do it are complex too
I wouldn't actually mind android-like permission model for out-of-distro packages (snap/appimage/etc.), maybe a bit expanded so I could say set this this and that folder for the "graphics editng app", and maybe save that as a profile to apply to some other similar app to ease on repetition/alert fatigue.
With maybe a layer to abstract some operations to not be just "allow this(remember choice)". Like file opening, if app calls to open a file I "just" want DM/WM specific open dialogue, with app/container name in the title and select the file to open.
Same for editing, I'd want to be able to just get dialogue "open file for editing", with app name and the permission to edit said file saved for the duration of the session so app doesn't need to re-ask me every time it saves the file.
Indeed. The idea of flatpack is to change desktop linux culture by normalizing the installation of 3rd party software, particularly proprietary software that people otherwise wouldn't trust without some form of sandboxing.
Who does this benefit? I can think of two groups of people. 1. Commercial software vendors who want more Linux users to install their proprietary software. 2. 'Transplants', new Linux users who are already accustomed to the Windows/MacOS style of wantonly installing proprietary third party software they downloaded off random corners of the net, and don't want or know to change their habits.
The value proposition for experienced linux users who don't do that sort of thing in the first place is next to nil. The only applications that might benefit from such sandboxing are applications like browsers, which have large attack surfaces and might be compromised while browsing the net. But even this is mostly theoretical, not a realistic day-to-day concern for typical linux desktop users.
> The value proposition for experienced linux users who don't do that sort of thing in the first place is next to nil. The only applications that might benefit from such sandboxing are applications like browsers, which have large attack surfaces and might be compromised while browsing the net. But even this is mostly theoretical, not a realistic day-to-day concern for typical linux desktop users.
You are jumping to conclusions here. RCEs are probably more common than you think, and I'd prefer anything that interacts with the Internet to be sandboxed.
Flatpak allows me to easily sandbox Steam games. It provides an easy target to tell user to test against to eliminate distro-specific issues. It allows to run glibc-only software on distributions such as Alpine. It allows me to have multiple versions of a program installed concurrently. It prevents programs from cluttering my home directory, and sandboxing gives me extra peace of mind. As a non-root user, I can also install flatpaks. Ostree also usually makes updates more efficient.
If you use a couple flatpak apps, they are available regardless of your distribution. That helps when working on multiple different distributions.
Use an old-ish debian but need a feature from the latest unstable software ABC? Install ABC as a flatpak, and do not compromise the stability of the base system by enabling all sorts of external, unstable sources.
In those 6 years, how many programs packaged and distributed by Debian were exploiting that?
If you can run the "beep" command, you can also edit the user's environment and from their easily escalate to root anyway. In modern desktop linux, the user is almost always the admin as well, a single person using their personal computer, so getting root is merely a matter of waiting until the next time that user uses sudo/etc. Windows tries to mitigate this sort of attack using secure UAC prompts that are apparently difficult for attackers to emulate, or so I've been lead to believe. But common desktop Linux distros don't require anything like that. Instead, the user has to be cognizant of such possibilities and not run programs from people and organizations they don't trust.
Yeah, because I definitely double check the provenance of the 30 dependencies that blow past my terminal when I apt install something, that I also very much looked into and aren’t blindly typing commands from Stack Overflow into my terminal because I’m trying to solve some problem.
because I definitely double check the provenance of the 30 dependencies that blow past my terminal when I apt install something
why would you? that's the package maintainers job.
each of these dependency also has a maintainer, so by definition all dependencies have a provenance that is as good as the package you are installing.
this is not npm where anyone can upload something and you have to check the provenance of each yourself
Repos are generally safer yes, but they can still act as vectors for malware.
There’s also systems like Arch’s AUR which is quite popular and more likely (if still unlikely) to carry malware, to the point that the Arch Wiki warns that use of AUR is at the user’s own risk.
Plus, many people are going to need to use proprietary software, which is always an unknown and likely to act badly in any number of ways. A lot of such software is Electron-based to boot, and devs are notorious for using ancient (and vulnerable) Electron versions.
every package system (apt/yum/pkg/whatever) is distributing binaries.
So yeah, the upstream project can be open source, but there is 0 guarantees that the binary I install on my system is the exact same binary as I would get if I build the source myself (and this does not even touch on the subject that compilers can add weird stuff as well)
Sure, it's better than closed source, because at least you have the possibility to check all this. In practice though, we outsource this responsibility to the package maintainer of the package system we use.
there is 0 guarantees that the binary I install on my system is the exact same binary as I would get if I build the source myself
not true, for years there are efforts in various distributions to make package builds reproducible. there are ways to build a package from source that allows you to get the same results and verify it.
we outsource this responsibility to the package maintainer
which is the point. i trust the package maintainers to do a better job at that than myself.
Escape hatches should exist, but I think it's better if those exist on a per-program basis.
If a systemwide "disable all safeguards, give all programs access to everything all the time" switch exists, the level of friction encountered when accessing it should be very high to help deter social engineering attacks. It's a one-time action so the annoyance level of that friction is negligible, since those using it will only need to do so on clean installs.
This why there are settings for this sort of things. Nobody in this thread is saying that something was impossible, just that some settings had to be changed and the UX was suboptimal.
These permission systems in practice don't really do as much to shield users as many think though.
People often just drop the word “sandbox” and say “applications are sandboxed” and that that means that they're safe but it's really not that simple in practice. What often happens is that such applications still need to communicate over some socket with some server that was never designed for such a sandbox, say PulseAudio, and in many cases can then simply instruct the outer daemon to do whatever they want with full permission, either by design, or by oversight since the no one who wrote the outer daemon thought about it at the time since they were never designed for that purpose.
Of course, but that just means that said daemons need to be reworked to not have access to everything either.
This is why there's a push to do as much as feasibly possible in userland in both macOS and Linux, so even when a bad actor tries to route through system components the blast radius is limited. Realistically, they should be sandboxed too — an audio daemon for instance has no business directly accessing storage or network facilities for example.
>> I think permission systems are bound to wind up in all desktop operating systems, eventually
What I'm about to say may seem wrong, stupid, or crazy at first. I think permissions often belong in the GUI. Applications would get no access to the file system directly, but they could use an API in the gui to open files - only files that are granted access by the user, often by selection in a File->Open dialog or other direct user interaction. By putting the granting of access in the GUI toolkit, we can run untrusted apps natively with no OS permissions.
Maybe not directly in the GUI, but something like that. Trust the user but not the app.
At some point you have to trust the user to choose the apps they want to run. That's simply not the OS's job.
To the extent that it is the OS's job, you don't have a computer anymore. You have an appliance. Sometimes that's OK; I don't complain because I can't run Doom on my dishwasher. But let's be clear about what is a general-purpose personal computer, and what is not.
> To the extent that it is the OS's job, you dn't have a computer anymore. You have an appliance.
In my opinion, that depends on the existence of an escape hatch. If it's like iOS where there effectively is none, sure, but if it's like macOS where SIP, Gatekeeper, etc can be temporarily disabled to make changes and then re-enabled or disabled entirely it's a different story.
An admin password prompt is hardly a deterrence to people doing stupid things. A young physics PhD friend of mine fell victim to a tech support scam, happily installing whatever spyware “Apple Support” told her to install over phone. That was a few years ago. The average person is too easily social engineered into allowing anything.
Sure, I don't think either this[1] commenter or Ken Thompson were trying to say that the product category shouldn't exist. A computer is vastly overpowered for what the average user is capable of or interested in doing[2], which is why toy devices like iPads are so popular.
I interpreted both of their comments as claiming that the direction MacOS is taking is a poor fit for those who still get value from powerful, general-purpose computers (myself very much included! I occasionally have the misfortune of using Macs, but am much much happier on systems where I can dig as deep into its layers as I need to solve my problems or scratch my itches)
[2] Though I do think it's a minor tragedy that the increasing amount of guardrails has narrowed the opportunity for an inquisitive youngster to explore his computer's internals
> The average person is too easily social engineered into allowing anything.
How many "average" users you know who use sudo? At some point, the software needs to acknowledge users who are saying "I know what I'm doing and the risks, just let me do it" i.e. sudo.
An admin check tells the OS that you are an admin, not that you know what the software does and that you are ok with CoolWallpapers logging all inputs.
So what should happen when the threat model changes? Just abandon all software, ossify it in a poor state, or something else?
You always to be advocating for ossification to avoid breaking apps which are no longer ok under an evolved threat model.
Finally, you didn’t actually answer the question I asked. It’s all very well and good to say how things should be, but people have to face the world as it actually is instead.
If applications can edit arbitrary files on the system it's already game over. I have no idea why people focus so much on “keylogging” as the supposed super important and dangerous thing.
If one run any malware with the full file edit permissions of one's user account at that point in theory the only solution is erase not only the hard drive, but also every other drive on any other system one's user account has access to or at least in sofar those do not have some logging for connexions in some way to see who connected that cannot be edited by the permissions one has on that system. Of course if one has root on one's own system nothing on that system can be trusted any more from that point. The malware could in theory have edited the firmware at that point to hide any checks one could do with a recovery system on a portable drive, but that's all quite theoretical of course, but it's possible in theory.
Keylogging is such a strange thing to focus on in the face of being able to edit arbitrary files owned by the user.
It doesn't matter and it's still a theatre. Those malicious applications can do what they want regardless by editing arbitrary files and obtain the same end.
The supposed threads of malicious applications keylogging and stealing your website passwords to worry about is rather strange when such an application can edit the files on your system such that you're starting a modified version of a web browser they injected with whatever code they want to do the same. In fact, this is probably easier to do than try to write some kind of a.i. that filters what it thinks are “password keypresses” opposed to altering the code of the web browser such that it simply sends whatever is being put into a field marked as “password” on a website.
It's a moral panic boogeyman that has no actual implications for actual real life security. Like quite a bit of “security” talk these days. Much of it comes down to the “door in your room” analogy where “security experts” talk about putting a big door in the middle of one's living room with an impenetrable lock on the idea of kindly asking criminals to only go through that door to steal things. In reality they'll just walk around it, and now one has an inconvenient door in the middle of one's living room.
> By 2019, you had to enable the application explicitly to listen for events at least in three places deep down in the system preferences, click accept in various popups and if you stuck somewhere then nobody could tell why it wasn't working.
It’s an improvement for users because it means that not all random applications and programs that run can act as keyloggers. It’s optimising for the common use case (random people running random software and being very annoyed if they get ransomed) against the rare case. It’s the same thing with debuggers and attaching to other processes. In the end it is a good thing to not be able to do that without explicit authorisation.
> So I had a very expensive laptop and the OS didn't let me use it freely.
It does not prevent you from doing it. It added some friction, sure, and you can find that this friction is unacceptable (and changing OS every now and then is a good idea in general anyway). But from a fundamental perspective the functionality is still there. The OS still lets you do this.
> I created the best looking/most usable desktop experience MacOS will never have.
It is great that you have both the ability and the time to do this. I’ll look into it for my Linux boxes.
However, my experience is that it’s never actually “the most beautiful/user-friendly/consistent/polished” (things we see all the time with new DEs). They all tend to fall apart with millions of corner cases and inconsistencies every time you get off the beaten path. In any case, good luck with your project.
Sure optimizing for security makes sense, but Apple isn't just doing that. They're also removing ways to override those restrictions. Often old methods to disable them or to whitelist an app silently stop working. Sometimes new ones don't always work, or require an absurd number of hops.
It seems alongside security there appears to be a strong desire at Apple to make macos a walled garden like iOS devices. They've hamstrung mobile safari for years to ensure the app store doesn't have competition from web apps.
Given that it should be difficult, if not impossible, for random applications to listen in on user actions, is there a better way Apple could have done this?
A popup on the first start with an alert and a password prompt is a good solution, but usually too many users type in their passwords blindly in case of a password prompt, so I would go with just an alert with "you need root privileges to run this app" and then they have to figure out 1. why do they need root privileges 2. how do they start an app with root privileges.
The security barriers for getting keyboard events in macOS are not as simple as “root has access to everything”, and for various reasons can’t and shouldn’t be that simple. So ignoring that part, Apple could make this into a one-click solution, but the number of legit apps that need to do this are so small that it’s very unlikely they will dedicate engineering time to it.
This is obviously where open source is superior. Apple probably can’t justify cleaning this up in macOS, but you can just go in and make it easier on on Linux if you have the knowledge and time.
But at any rate, Wayland is completely optional. You can keep running X and nobody will stop you. People will keep running X without issue for a very long time. This is very different from what the Apple world is like.
Many of the things one can at least ask permissions for as I read it on the Apple system are actually purposefully simply not available due to similar concerns.
Many of the leading Wayland developers believe in this. The Enlightenment lead developer for instance does not believe a programmatic way to make screenshots or listen to keypresses should ever exist. I had some debates with him about this and he believes the risks are too high as well as revealing that he seemed to believe that streaming videogames did not work by way of a third party tool that captures the screen contents, but that each video game had this functionality built in, which is certainly not the case and that he believes this might explain his reluctance for such an a.p.i. to facilitate this.
> But at any rate, Wayland is completely optional. You can keep running X and nobody will stop you. People will keep running X without issue for a very long time. This is very different from what the Apple world is like.
Opinions are divided on that matter. The reality is that many of the developers of Xorg have abandoned in in lieu of Wayland and many Wayland developers, many former Xorg developers are clear in their opinion that they see it as a replacement, not an alternative and eventually expect everyone to switch.
Whether that will happen is anyone's guess. They are often met with counter arguments that Xorg and the X11 protocol itself simply has features that many businesses and private individuals need for their lifelihood so there is going to be commercial incentive to pick up maintainership should they abandon it. They have also conceded heavily already on many of the features they initially said where either unneeded or a security risk when they realized the reality that many people outside of their bubble did use them. Libinput originally did not have any mouse acceleration settings on the argument that no one would want to turn it off or fine tune it to begin with, but now has it when they realized that unlike what they thought, demand for the ability to fine tune or turn off mouse acceleration is higher than they anticipated.
>> People will keep running X without issue for a very long time
> Opinions are divided on that matter.
I feel like those holding the opposite opinion are pretty clearly wrong and it seems to be some kind of hubris. Fortunately their opinion is not binding.
> The reality is that many of the developers of Xorg have abandoned in in lieu of Wayland and many Wayland developers,
Not an issue. It doesn't need a lot of maintenance. It's "done". Source is available too. Just needs minimal effort to keep the lights on. As long as someone is interested in running it, as many will be, it will happen.
That is a very wrong view of something that communicates so closely with the hardware and deals with graphics.
It's very much chasing a moving target of new graphics technology and many of the innovations made by Wayland have been ported to Xorg. The interesting thing is that screen tearing used to be a problem on Xorg but never was on Wayland but much of the technology responsible for that has now found it's way into Xorg to make it a thing of the past there too. Of course integration with new hardware drivers is also required.
It's entirely possible it will continue to be maintained that way, or that the Wayland philosophy will finally be broken and these things will be added once they realize there is a great enough demand, but these are all possibilities, not certainties.
I don't think there are specific issues. The problem is that with each release of macOS usability, reliability and performance regress. If you're interested in these things and not so much in "bling" and questionable new features, then you'd share probably Ken's "atrocious" opinion.
An example is the latest System Preferences. It's virtually unusable.
Seems overblown to me, been a macos user for over a decade and the ios-ification o f the mac hasn't been a big deal for me. My greybeard workflows still work the same (everything in iterm, using vim, tmux, all the "standard stuff"), sometimes updating xcode is a pain but w/e. Homebrew still chugging along. Parallels is fantastic. Performance of the M1 Max on my new mac studio is just heavenly. The changes to macos have been nowhere near as egregious as what MS is butchering windows with.
The latest update added "Stage Manager" or whatever, tried it out for an hour...didn't really like it and turned it off, it never gets in my way trying to force itself on my like Windows does with all of its anti-features.
People complain about system preferences but I use spotlight to find settings when I need them, which works great - and once a setting is set I rarely change it. I don't think I've touched the preferences "app" since I initially setup my mac studio. The moaning and groaning about how bad it is just seems so.....pointless to me.
Been using Macs exclusively, since the iPhone came out. At some point I noticed a "do not disturb" icon displayed on the top UI bar, something that somehow got turned on after upgrading to Ventura. It took over half an hour of googling and trying random shit before I was able to turn it off. Yes, Settings is a UI disaster, even ignoring that...
Goofy shit like this randomly happens, particularly after OS upgrades. I remember some crazy hell where I couldn't finish the OS install on a new Mac, because Apple decided that Apple IDs must be email addresses, and my ancient ID was not. Required a call to customer support.
And the "untrusted developer" shit always bites me in the ass every couple months or so... It is particularly painful when using pro audio interfaces, which will just suddenly stop working every now and then. It requires a magic key-press during a reboot to clear some kind of special RAM.
Except that spotlight doesn't find all system settings. And I've already ran into a bug where the system settings dialog loses sync with the configuration requiring the app to restart. And also had settings pane just turn blank.
It’s fine if search is a way to use System Preferences. The problem I have is that since they turned it into a “big list” search is essentially the only way I can use it. They destroyed the spatiality of the app that had created grooves in my brain over the last 20 years. Now it is very disorienting to use because I can’t efficiently browse for things.
I love comments like these. Pure macOS Stockholm Syndrome. “Sometimes some things are a pain but it’s pointless to groan about it.” I mean you do you, but don’t blame other people for finding some of these issues more than they can stand.
> I don't think I've touched the preferences "app" since I initially setup my mac studio. The moaning and groaning about how bad it is just seems so.....pointless to me.
Have you considered that other people have usage requirements different from yours?
Of course there's not much reason to complain about System Settings if you never use it, but that's missing the point.
His usage requirements is using it only once when he sets up his computer, which is probably a good thing for a preferences app. It works well for him, why shouldn't be able to say so? So we can keep this thread strictly about complaining?
You could copy paste your comment to everyone here, "have you considered that other people use it differently".
> It works well for him, why shouldn't be able to say so?
That's not all the commenter said. "The moaning and groaning about how bad it is just seems so.....pointless to me."
If System Settings is fine for the commenter, that's great for the commenter. (Although "I virtually never use it" isn't exactly a great response to "It's virtually unusable" or a great defense of System Settings.) However, the commenter is criticizing other people for complaining about it, and that's not justified.
He undermines his own point. He says he has no problem with System Preferences, and then essentially says he never uses it, explaining why he would have no problem with it.
I'm sure you already know it doesn't contain spotlight itself, which belongs to Apple.
There exist equivalents that have exactly that same functionality as described above. I don't know what GNOME uses, but KDE's krunner searches system settings, user files and applications, bookmarks, browser tabs, solves arithmetic and unit conversions, does spell checking, searches the web, searches active windows, applications and workspaces, etc.. It's configurable with many plugins available. I expect GNOME's system works similarly.
I'm specifically speaking of Ubuntu's version of Gnome, I guess. Press the Super key and a search box appears, a la MacOS's Spotlight. This is how I open apps. There are likely nuances about what is searchable, but in practice it feels the same to me.
Does Homebrew still have those issues on newer Macs with the /opt/homebrew prefix? I don’t use either at the moment but last time I looked at Homebrew it looked like they had changed the way it installed, at least on Apple silicon Macs.
I'm using Nix now on my macOS installs, so I might not be totally correct the current situation. But yeah, I believe most of those sorts of issues have gone away with the new prefix.
There's still some weird annoyances for me, though. For example, it's still intended for single-user mode only. The best solution I came up with was to create a separate user for Homebrew and then basically alias `brew` to `sudo -H -u homebrew brew`.
And generally, if you attempt to use a non-standard prefix, such as in your home directory, packages will have to be built from source. I understand why, but it sucks because this means when you need x86-only packages via Rosetta, you're stuck with the old `/usr/bin` prefix, unless you want to and can build from source.
Also, in general, maintaining multiple prefixes/architectures is annoying too. I wish the default command allowed me to just pass `--arch amd64/arm64` or something when installing packages
The default install location for Apple Silicon and Intel is different, for some reason. I got weird problems when many bottles were still Intel-only, but now everything is good. So unless you hate /opt/homebrew, it’s great!
The latest system preferences are an improvement to me. It was a mess before and it’s still a mess but the app layout makes it easier to browse, and at least it’s consistent with iOS.
This obsession with “unifying” disparate things under One True API / UI / Language is all over software. It’s probably the same mentality behind what what Microsoft was doing with Windows 8 and Metro. Unnecessary abstraction is the bad Software Engineer’s bread and butter. We’ve all worked with That Guy who kept arguing “We could call X and Y the same way if we just had a unifying abstraction Z. It would simplify everything!” Z becomes the worst of both X and Y and the customer is less satisfied just wants to keep X and Y separate. But the developer is more comfortable now, and that’s what matters.
Everything seems to get sacrificed at the altar of Developer Comfort. Performance gets thrown under the bus so we can write everything in JavaScript. Platform-specific features get abandoned or neutered in a cross-platform framework so we only have to target one API. We ship gigs of Docker containers so we don’t have to get our software to work on all the customers’ computers.
This is just another example. Wouldn’t it be great if iOS apps and Mac applications converged on the same thing? Only a developer would want this.
Right? It’s a middle finger to mac users with an external screen. It was too hard to make it scale horizontally, or even prefetch the settings in the background and not have a lag between clicks. Not the end of the world, but backwards
The very argument that was presented when iOS first came out, as well as whenever touchscreens on MacBooks come up, is that the UX is fundamentally different.
So why would it suddenly be a good idea to import this from iOS?
Perhaps the people against touchscreen on MacOS finally retired or gave up the fight, so Apple is slowly preparing what Microsoft did years ago with Windows 8.
What Microsoft did, and still hasn’t finished. System preferences still contain massive amount of details contained in Windows 7 (and older!) style windows.
At least Apple managed to move all their settings from one paradigm to the next in one year. I’m looking forward to the improvements they make to the app in the future.
They didn't finish it because they went hard, all in one go, and suffered consumer backlash as a result. From this Apple learned that you have to boil the frog slowly.
Because it's confusing to have two ways of doing things when one way works just fine.
Why would you ever want added cognitive complexity?
If things need to be different then sure. But if they can be unified in a way that works great for both then please do!
Honestly, with their tight integration, I don't even think of watchOS and iOS and macOS as being so separate. They're all just a kind of "appleOS" that gets applied to different form factors. So UX unification is ideal wherever it makes sense.
And thanks to this peculiar way of thinking we also have several layers of UI inconsistencies in Windows, for example. Someone thought "let's unify everything" but, assuming this is a good idea at all, you really need a lot of work, planning, effort, and most of all imagination to make sure the final result is actually an improvement for the users of both mobile and desktop systems rather than being a mediocre one-size-fits-all solution.
Those are the first system preferences on macOS that are usable to me, and I'm not even a big iOS fan. The chaotic version before felt like Windows to me.
Say what you will about the new version, but the old version of System Preferences was certainly not "chaotic". I found it very pleasant to look at and navigate.
This new thing violates much of Apple's own HIG, and despite running Ventura for half a year, I still find myself hunting for things that should be easy to find. Quick: where do you go to turn the volume menubar control on or off?
I agree. In the old SP I usually used spotlight to find what I wanted as quickly as possible. Now, I usually still do. The practical difference is negligible. I have no trouble finding the settings I want to modify, and a rarely have a need to modify any.
My response is similar for the other complaints I see here. I can ignore new features I don't like. I've got integration with my phone and watch that took zero effort. No need to waste time on drivers, etc. If I want to run Photoshop I can.
(I've used Macs since the original Mac Plus with one one-year Windows interlude. For servers I use Linux, which is obviously appropriate for that use.)
As time goes in from 10.5 System preferences UI kept on degrading to the point that I didn't know what icon to click and frequently just made the search bar for everything. Having the search bar more prominent and making the UI standard for all of the other products probably is a good idea.
What code is prevented by security? Other than limits to kexts, your code isn’t beholden to most of the system security changes unless you specifically opt in to make a sandboxed app.
I have to deal with refusal to install apps from "untrusted developers" constantly, I can't even believe you wrote that.
For starters, the UI experience is terrible. You get a dialog telling you you can't install an app; close the dialog then you have to know to go to a certain Settings tab under Privacy & Security, and there will (hopefully) be some text that allows you to enable the app. This is a UI disaster. Maybe they do it so that non-technical users will give up in frustration.
Once MacOS decided that my official Oracle JRE was "untrusted", and I would get the bullshit dialog every time I started a Java process (note: I develop apps using the JVM). I had to google to learn some arcane CLI magic to disable the untrusted bits on my JRE files.
More recently, I couldn't get CIV 6 to run. Instead of telling you what to do, you get a "app is corrupted" dialog (maybe this was the fault of Steam). This required multiple enabling via the magic permissions tab. I mean, are Steam and Firaxis not "trusted developers".
All this pales in comparison to the pain when using my MOTU audio interface. Getting it to work was an enormous pain in the ass thanks to Apple security. And then MacOS would randomly decide to break things every 6-12 months, and getting it to work again requires discovering an arcane NVRAM reset procedure using magic key presses during reboot.
Your app is not beholden to sandboxing unless you specifically are using sandboxing as part of your application configuration. It would also only apply to app bundles.
I can make a command line tool or a standalone Qt application right now without having sandboxing pop up at all.
Even for access that the OS protects the user from, as an app developer I rarely need to think about it. When I try and access a limited area it asks the user for me.
Source: self after working on lots of varied code bases from web dev to 3D libraries and standalone applications for over a decade.
As an Apple user since the early days of OS X, I remember watching that part of the presentation thinking to myself "he knows he doesn't need to elaborate on that."
I think many of us watching Apple for the past two decades have seen the OS move slowly towards closed standards and tighter control instead of openness and functionality. Each time I boot my PowerBook G4 running Leopard for a nostalgia kick, I'm reminded of how great OS X once was.
But as Ken pointed out, they've been doing it over a long period of time, so it's OK (i.e., it's given us plenty of time to move to other alternatives).
If I had to guess, it would be Gatekeeper and SIP. These technologies can really get in your way if you’re used to having full access to your own machine.
Most people won't. And if you're developing software distributed to others, this will put your machine in a state where you aren't testing what your end users get.
And in that context, it’s better to at least have the option, should you ever feel the need to use it.
Personally I leave both on because even having used computer for 27 years at this point, I can still slip up and I’m still vulnerable to social engineering among other things, and so it’s nice to have something to help cover for those situations. It’s no cure-all, but it at least raises the bar for malware and such.
To me it appears Apple has lost all the focus after Steve Jobs, there is no "do more with less" spirit in Apple products anymore. They still got some Mac OS X inertia, but it's mostly about changing colors, animations, rearranging items and trying to address every possible use case - instead of offering new OS paradigms. It's interesting that their hardware division keeps innovating - people seem to love their M1s. I miss coherent Unix environment on good hardware. Snow Leopard times were great.
What bothers me is how they're charging developers to build for their system and semi-require server verification for programs to run.
In the earlier Mac OS X era they used to be very open (and free) with their tools and actively encouraging.
It feels very backwards now. For example, Microsoft was really restrictive with its expensive tools at that time and gradually opened them up. It's like the two companies reversed positions. As a kid, I could never afford to buy pro dev tools, and while it's not as expensive currently, I think that's the way Apple is headed. It's not going to help more people get into programming.
I recently tried to install gdb on a Mac. Well after 1 hour I gave up and went a different direction. gatekeeper was refusing to install code that wasn't signed out the wazoo and the 14-step instructions on how to sign, approve, bless, and sprinkle holy-water on top of a gnu gdb binary was just too much. it didn't matter what i did - even if i downloaded the source code, compiled and ran it as root, that wasn't going to make gatekeeper happy. gdb was never going to start.
if you watch this talk you can tell that gatekeeper is taking away your freedoms.
they are trying to turn each and every binary on your machine into a walled garden.
this will put apple in a superior position - the only company able to make a binary that can use every resource in the OS - and will put developers at a great disadvantage!
Isn't it obvious? Look at how Apple devices are used and look at what he built. I'm surprised he stuck around this long. He doesn't care if the touchpad is marginally better or their silicon is marginally faster because they are trying to destroy his world.
I don't know what his specific issues are but I've always found it odd how many IT professionals prefer Apple. The interface is so frustrating to me and there is so much the OS makes hard to do. It really does get in the way a lot more.
Apple addresses a market of people who entered the market over 5 decades. Most of them way after Ken. It went from improving on the Mac to completing the Mac’s destiny as a computing appliance.
I think apple is degrading general purpose computing.
It seems like you have to buy software to do anything on macos more conveniently.
Why can't I write a script in python? or a gui script in python? it's the top language.
You can use say swift, but even with "oh we opened it up", it's really an apple-specific language, and it's compiled.
Yes you can get brew going, but that's not apple.
and ios - what a travesty. You don't own your phone. You can't access your filesystem. you have to ask permission to do anything (and they don't grant it for most things)
sorry, yes, I know. they shipped 2.7 for years, but it's not a "first-class language". you can kind of do gui with tk. There's really no scripting language (applescript isn't really what I would call a scripting language, and really just not that good).
I feel the same way, but I know that moving to Linux isn't for me. My approach has been to run a Hackintosh, which provides excellent compatibility, reliability and parts availability, and resist upgrading macOS at every major update. I'm currently still on 10.14 and haven't been forced off it yet, and maybe never will. Generally, if something requires a more recent version, I just reject it. There tends to be a positive correlation between the quality of software and the age of minimum OS requirements anyway.
An interesting thing to note is that even with software that states support for an older version of the OS, it usually is more buggy, so there is a tendency for software to degrade over time as it gets updated.
That's odd - I'd prefer the opposite tradeoff. Apple hardware is excellent but for me their software design choices are usually the problem.
Even the much-lauded excellence in UX doesn't really hold up any more.
But at least you can close the lid on a Macbook, put it in your bag and be secure in the knowledge it won't decide to switch on and probably cause permanent heat damage to itself. (looking at you windows...)
No that's not it. The issue is actually pretty involved, there is a LTT video on it, this is the bit I'm remembering: https://youtu.be/OHKKcd3sx2c?t=475
Ah I see... you can disable network waking, also the ability of Windows Update to wake the computer. That, combined with disabling wake in other areas, solves it on various Windows computers I've tried it on.
I guess it's possible that all of them were defaulting to S3 sleep though, I never checked.
In Ventura (13.2), you can turn off automatic updates in the “General” tab, then “Software Update” and click on the little “(i)” icon to the right of “Automatic updates”.
When I saw it the first time, it was unclear to me that I can even click on it, since it doesn’t look much like a button. They shouldn’t hide such important settings in a tiny icon or at least should have made it much more apparent that it is clickable.
Couldn’t find any settings related to the closed lid though.
Thanks for pointing to that. I don't think I've ever checked this setting, but I just did and in the latest macOS it's located in System Settings>Battery>Options>Wake for network access>Only on Power Adapter and it was already set so I'm good. Hmmm. I don't remember ever changing it from the default.
I have to admit, I tried and failed to switch away from Apple. I bought a Linux Laptop - an System 76. I was surprised at how terrible the battery management of linux was compared with my mac. And that particular issue broke me.
There simply isn’t a proper contender to Apple. And with Apple silicon, it’s a done deal for the next decade at least.
Apart from hardware, you need a company that is willing to take Linux from the ground up and create a macOS type OS. Not simply make your own distribution and DE and call it a day. Chromebook was close, but they had bad execution and wrong ideas.
Linux is a kernel. Making your own distribution and DE is how you create a Linux-based macOS.
I’d say System 76 is doing that but their execution has stumbled for the past year or so. They are working on their own Rust-based DE, to some level of success. I hope they get back on track.
How do I get full-disk encryption on Linux without having to type a password twice on boot? This is not a surface-level issue you can solve at distribution / DE level, and it’s not the only one.
A good pre-installed environment of current tools would be nice, of course.
For desktop, I had exactly the opposite experience in ~2006. I had a whitebox *nix box that had a catastrophic failure and I needed to be up and running quickly. I was working 12+ hours a day, had a new baby, and I didn't have time to install / maintain a *nix box. So I bought an iMac to use as a desktop. What a mistake.
Everything that was easy in *nix was a massive PITA in MacOS. I'm talking about things like basic desktop customization (focus follows mouse, customized mouse buttons + keyboard hotkeys combos to move/iconify/resize windows, etc). I ended up using open source X11 based utils for most things but my web browser & mail because as long as I stayed within X11, I could satisfy my 20+ year old muscle memory with focus-follows-mouse and my hotkeys. However, every now and then I'd blindly start typing and the focus was still on my mail client or browser and random things would happen (adding bookmarks, deleting emails, etc) After a year I never managed to unlearn ffm & my hotkeys, so I gave up, gave the iMac to my then in-laws, and built a new whitebox.
I realize there are extensions for ffm, but ffm on MacOS is a crapshoot in my experience and I was never able to find an extension that I like. Similarly, at the time, I could not find any extension that satisfied my muscle-memory window management hotkey/mouse button combos. I don't mind click to focus so much on a laptop, but I can't stand it on a desktop.
Over the years I have settled on *nix on a desktop and MacOS on a laptop..
It requires lots of manual control but from my experience I can match or exceed battery life in comparison to running windows, with some tweaking. Out of the box it was terrible as well.
And this continues to be why Linux suffers with the mass market.
I ran Linux solely from 1997 to 2004 on the desktop and switched back because it was such a pain in the ass to manage. Every time I go back, it just isn’t catching up to the mainstream OSs’ ability to manage without “futzing”.
My M2 currently lasts all day on the battery while spending half of it on conference calls and half of it writing code. Nothing can match that right now, but I am still hopeful for a Linux future.
I get irrationally angry when people imply that Mac and Linux are somehow similar, or interchangeable.
They are not! Not even close!
I think this is where a lot of the negative opinion of Linux comes from, people lament some issue they have with their Mac and someone else will say, "oh you should try Linux, it's really good now." And so they do and are surprised to find that experience is vastly different.
Linux is a great choice for people who want the highest degree of control and freedom over how they interact with their computer. The trade-off is that sometimes you have to futz with it or report bugs.
If you view your computer strictly as a tool to let you get other things done, then you want Mac. All of the OS and UX decisions have been made for you and you get what you get, but you (should) never have to tinker with the system itself.
I use Linux because I don't want to futz with all the crap in MacOS and Windows. Takes maybe twenty minutes to install Debian and then I touch pretty much nothing for years, except maybe change apt channel if I want something updated faster.
Everyone I know that is stuck in big corporate OS spend more time than that every week fighting something their masters push on them.
Application install is a terminal command that's trivially easy to learn, there's no e-shop solution with ads and surveillance, and while I use i3wm I think most people would be about as comfortable in XFCE as in the MICROS~1 OS:es of yesteryears. You decide when to upgrade, there are no nagscreens or forced reboots.
At least for the last five years or so I've had no trouble with UEFI, WLAN or sound.
Fair enough; if you’re talking about random machines Linux light be installed on.
But! I believe System76 should ship their laptops with tools and/or settings and/or configs for other mainline distributions that maximize the experience on their hardware while allowing for the inherent customization Linux-based distributions.
I’d argue Sys76 is far closer to Apple than just installing Linux on a random machine.
I prefer Linux (fedora with gnome) because it needs less tinkering out the box.
It's the basic things like not having to install third party utilities to have window centric window management (as opposed to app centric window management). Or being able to plug my Android phone in and be able to browse the files without additional utilities.
I disagree that it’s a trade-off. Linux could match Apple’s out of the box battery performance. It’s just that nobody seems to care about Linux’s “out of the box” default configuration because, well, the user can futz with it! Things are slowly getting better though. But we still don’t have that distribution where things are already futzed for you and everything runs as well as it can.
But if ou bought a Linux laptop, it should come preconfigured with the optimal settings, with at least one or two distros like Ubuntu and Fedora, plus the settings described in a manual.
Similar experience over here on one of the earlier 11th gen models and recently on a new 12th gen model as well (build quality feels a lot improved vs the early stuff too, fyi). People are still right to call BS on having to run any of this to get decent battery performance but it is possible to get if you jump through a few hoops. Next step for distros should really be sane defaults for this stuff and/or an option to tell the installer it's going on a laptop/non workstation desktop so that it gets sane defaults for the laptop based on efficiency per watt rather then the sane defaults for a server/workstation (max performance).
I can understand distributions not having optimal settings for every laptop under the sun, but it seems like a no-brainer to at least create a repository of power config profiles for different models that users can submit to, eventually creating a decent library that could then be integrated into the settings app or and first run wizard.
Yep, and another frustration with Fedora in particular are the defaults for DNF. I don't get why the distro doesn't set `/etc/dnf/dnf.conf` with `max_parallel_downloads=10` and `fastestmirror=True`. sudo dnf update is so much faster with that (though you still need to keep away from the GNOME software app).
I havent really ised redhat based distros since redhat 5.4....If you want to go for a walk on the wild side try cachyos. It makes arch easy and has some nice defaults (atleast for kde, not sure about gnome, as i went kde them recently hyprland). It also has v3 and some v4 architecture compiled packages which are more performant on newer cpus. The pacman defaults are reasonable as well regarding parallel downloads.
Two of my last five laptops were macs, three linux boxes, if you consider corporate laptops as "mine". One of the three linux laptops just worked beautifully (wrt. battery and other driver issues), one didn't at all, and one worked poorly until I installed some software that shouldn't matter for battery lifetime, and since then it's had excellent battery management. There was a long list of dependencies, so I suspect that the software I installed depended on a package that solved whatever the problem was.
You may chance to buy hardware where someone really has tested and fixed linux. Or you may get something else.
But generally powertop is enough for a significant boost. For certain manufacturers (ASUS is a good example), you want to use their vendor specific tools to manage CPU/GPU powerbands.
I get at least 5h on a Thinkpad T14s (Ryzen 6850U), without tweaking, using "balanced" mode. I know some people think 5h is not enough, but it's plenty for a full day of work. If the System 76 system was Intel based, that might have been the problem, AMD has been way ahead of them for power usage in the last few years.
Raw battery life aside, I think one big advantage macOS still has is that its power management isn't modal. There is a low power mode like iOS if you really want to eek out as much runtime as possible, but I've never once hit a need to use it.
Outside of that, you lose no performance when running on battery, you get full performance when you need it, and high efficiency when you don't. Needing to manually choose a power/performance profile feels incredibly archaic to me, though I appreciate that some folks may want that level of manual control.
Plugged-in performance while unplugged without torpedoing battery life is what is really special about M-series laptops for me. It’s such a stark contrast with my ThinkPad which becomes noticeably more sluggish the moment it’s untethered in “balanced” mode and eating through battery like candy with “Performance”.
I see your point, but I just leave it on "balanced" unless the battery is critically low. Fan very rarely comes on, everything runs quickly, and I get decent battery life. This was NOT, btw, my experience with Intel laptops.
I have a Thinkpad T495 which is also AMD based. When I 1st got it battery life was atrocious under Linux. Kernel release 5.17 greatly helped battery life. Now it is just horrible even with TLP installed. The battery life under Windows is about double what I get when running Linux. It isn't the only but is the main reason I run Windows and not Linux as my main os on it.
That's more than three generations older than the 6850U I'm using. It predates the M1 release. It's only a few years ago that we could expect decent battery life without tweaking on Linux, probably just after the T495 series. (-: These days, tweaking with TLP &c could yield worse battery life, in my experience.
i switched. I have the previous gen AMD powered (Ryzen 5700U) system 76 and the battery life is decent (6-7 hours).
I have one 4 year old system 76 notebook with the NVIDIA gpu and that thing when running the 3d graphics mode is terrible (<3 hours), but its a gaming laptop and a beast (well 4 years ago it was..). I have 2.5 TB of storage, huge memory. Its more of a portable desktop than a notebook.
I really like software development on Linux much better than mac. There are a few things I miss from the apple side, but generally its been great.
I feel like this comment has short memory. Battery life on intel macbooks was terrible ime. Frequently it would drain completely while the lid was closed overnight. Always had to keep it plugged in. Of course Apple silicon is fantastic and nothing beats it now.
On the Linux side, I've had great luck with AMD + recent kernels. I get around 6 hours now on my Thinkpad X13 gen 2 which is good enough for me.
Raspbian* apparently. That's really interesting; even the highest tier Raspberry Pi still feels pretty sluggish as a desktop thanks to the limitations of SD card throughput/latency/queue depth/etc. I wonder what his usage looks like.
With the pi 4 you can boot off a USB3 drive like a SSD. I setup a 4gb memory model with SSD boot and it's a flawless little desktop--it runs full gnome and Ubuntu just fine. If you can get a 8gb memory model it would be perfect. I love it.
An Intel NUC is not much more than the top tier Raspberry Pi, once you get a case and stuff it needs, but is orders of magnitude faster. I got one and installed Linux Mint on it. It's a much, much better system than the Raspberry Pi 4 I have.
I bought my pi 4 at launch in 2020 before the shortages and scalping, it was $49 IIRC. There is no Intel NUC that's only 50 bucks. And I like that the pi is aarch64 architecture as I am developing for that architecture.
I was using a Raspberry Pi v1 as my NAS for years. It was fast enough and got the job done to my satisfaction. However when that pi eventually died and I tried to replace it, I found that v4 Pis were 1) just as expensive as NUCs, and 2) not even available for purchase anywhere that I could find. So now I've got a NUC filling that role. The Raspberry Pi org has really dropped the ball.
That is unfortunate about the R Pi. I got a 3B+ a year before 2020 and it better last a long time.
I think the foundation’s pressures are similar to Apple’s pressure to upsell their paid service all over the UI. That’s what I resent about Apple; it’s enshittification like with EBay, Amazon and even Google bugging me to sign in to make searches.
It may depend on one's definition of "not much more". I got my Intel NUC in the barebones configuration since I already had spare 32GB of RAM and an SSD for $387.59. To me, that's not much more than the highest end Raspberry Pi 4 with a case with a heatsink and fan because the NUC: actually works and is usable, has DisplayPort and just generally better I/O aside from the Pi's GPIO, configurable memory, and better CPU and GPU. Plus, you can actually buy one. To me, that's worth it.
My Raspberry Pi's are just unusable for anything other than as high-level embedded platforms. I've started selling them off, only keeping one or two for embedded use cases.
When it comes to how sluggish a system feels due to its disk, it's much more useful to measure the read latency and throughput of random reads because that's what the system is doing: you read a lot of sectors randomly when you boot or start chromium.
Note that the sample size is just 1, so I wouldn't place value on any single benchmark (it could just be that they chose their USB disk and SD card poorly or it was too cloudy that day). This is evident in my URL's hdparm result being wildly different from yours.
And? For desktop use it doesn't matter. If you're not paging out of ram you will never ever notice. I'm not sitting here running disk benchmarks all day... I'm browsing some web pages and using a text editor.
> If you're not paging out of ram you will never ever notice.
What you say is true but you didn't respond to someone complaining about ram latencies, you responded to someone lamenting poor disk access perf. And you recommended booting off SSD which is demonstrably specious advice.
> recommended booting off SSD which is demonstrably specious advice
Booting the Pi 4 off a quality USB-connected SSD to address stability and performance concerns is really good - and normally not controversial - advice. You're betraying a lack of familiarity with the subject matter here. The Raspberry Pi people provide a whole forum where people can discuss this, and other, stuff and educate one another.
Again I will tell you--it doesn't matter to me and I suspect 99% of users. Unless you are like a Linus tech tips fanatic and eeking out every percent of performance for your Good Gaming Rig then it's fine.
4K might cause some problems in my experience, but 1080p should be totally fine.
(in my memory the problem was more with running the display itself in 4k and trying to get a full frame rate out of it, but I think the advice still holds - 1080p videos on a 1080p or 2K display shouldn't be any sort of major challenge)
Because not all SD cards are the same and I have more large SSD drives sitting around than SD cards. Why do you think results for one SD card map to all of them?
Again, it doesn't matter. Real people in the real world aren't performance tweakers.
Oh no! Someone on the internet suggested someone do something that isn't optimal to performance! Wow better go on a multi reply freak out about it!
It doesn't matter. I'm not losing sleep over an apt upgrade taking maybe 10 more seconds. I'd much rather live with that and have a little desktop that barely sips 5 watts of power total.
I measured apt upgrades in multiple minutes on the Pi 3B. And forget about doing anything else; the system would hang hard until I/O was cleared.
By contrast, upgrading my desktop tower from HDD to SSD was an incredible, dramatic speedup in terms of booting and especially apt upgrades. The latter became nearly instantaneous. Blink and you miss them.
Now, on the Pi I use only the heavy-duty brand-name sdcards. I have found that the weak ones tend to suffer badly from ESD. I am not sure how much lower performance is from the heavy-duty cards, but I doubt it is much slower than the fragile ones.
As a lazy developer I also dislike what’s slowly happening to macOS. Apple wants you to switch to sandboxed apps but they don't provide a way for you to do even half the things a traditional app can (because they can’t imagine them all up front). That’s just frustrating and lazy on their part and makes the developer UX shitty.
But, as an end user, what Apple is doing (bringing sandboxed apps and better security to the desktop) is inarguably the right thing to do. It’s a far superior position for the user and it greatly raises the bar for malware.
As developers, to me, it feels a little bit backwards. I guess my critique is that there must be a nuanced way to say “hey Apple you need to do a better job at supporting valid developer use cases” (and I’ll be the first to admit I have many grievances) while at the same time acknowledging that the increased complexity of modern computing systems is moving the needle meaningfully from a security standpoint and so we should be okay with having to work harder to keep our users secure. Like, I’d truly hope if we all switched to Linux, we’d find a way to make secure boot and code signing standard. Not just say “ah isn’t the old dying way of user-domain permissions nice let’s live here forever”.
Even Microsoft is pushing code signatures and sandboxed apps. We should be making a stink and pushing for these platforms to allow custom root signing keys and fully secure/sandboxed replacements for the functionality they’re taking away. Not just throwing up our hands and saying fuck security I’ll just use Linux. Not a great image…
Developers vs. users is a false dichotomy. Users need software. Without software, there's nothing to use. The native Mac software ecosystem is slowly dying. Eventually everything on the Mac will just be a cross-platform afterthought.
What do you mean by "need"? Increasingly, users are force-fed cross-platform web apps. I wouldn't say that's what they need, and you're not even saying it's a good thing.
If users don't need desktop software, then why do they need desktop hardware?
> Again, I’m not suggesting all this is good. But of course users _need_ local compute and native apps less than 10 years ago.
Isn't this a problem for the future of Mac, though? It seems to me that dumbing down the Mac is the opposite of what Apple should do. Why not place emphasis on what makes desktop special, rather than morphing the Mac into an overpriced iOS device that only runs web apps?
That’s an interesting point. I don’t necessarily think it will go that far for the record. But Mac is probably comparatively well suited to make a profit off fancy devices with low computing power.
> We have terms like DX and UX. It’s not a false dichotomy at all.
It's still a false dichotomy, because they're not necessarily opposed. Making the native Mac software DX worse can also make the UX worse, and making the DX better can make the UX better. When I said "Eventually everything on the Mac will just be a cross-platform afterthought", I meant that the UX is becoming totally crappy. We get Electron, we get Catalyst, we get iOS apps on macOS. Crap. Bad UX.
Lack of powerful software, due to excessive security restrictions, is also a bad UX. Endless "Cancel or Allow" dialogs are a bad UX.
Lazy in this context doesn't mean “lacks work ethic”. You might be unfamiliar with the Gates quote. Point is devs often whine when they have to take the scenic route even if it’s safer for the user.
Anyway, I’m not saying they’re at odds and you can only pick one. So we seem to agree. I’m saying devs should be applauding better security and not whining about the transition away from user-domain security which is kinda what is implied in the OP. We should be whining about not having access to to the system in ways we need. Better security needs to accommodate valid use cases not stifle creativity. We probably also agree there.
Opining Linux because it’s still the wild west of software and you can do shit like zip up the user’s browser history, logins, and cookies, or idk dump entire hardware profiles without any user interaction, or prompt the user to install a new boot loader or drop your rootkit into initrd because your program “just needs their password to update” is just that.
> I’m saying devs should be applauding better security and not whining about the transition away from user-domain security which is kinda what is implied in the OP.
It's not better security, it's security theater. Security researchers roll out of bed in the morning and find bypasses to macOS TCC (Transparency, Consent, and Control). I've done it myself, and I'm not a professional security researcher. The so-called "security" doesn't actually stop real malware, it only stops honest developers from making a living.
People discover vulns all the time. A vuln is not theatre. You patch the vuln and you are secure. A vlun does not security theatre make.
A fundamental design flaw in the security model or system which renders it irrelevant, on the other hand, is a problem. Are you saying there’s a fundamental flaw in Apple’s implementation of secure boot, code signatures, and sandboxing that makes it irrelevant?
Except you aren't, because there's an endless series of them.
> Are you saying there’s a fundamental flaw in Apple’s implementation of secure boot, code signatures, and sandboxing that makes it irrelevant?
I said TCC, and you said everything except TCC. (Although sandboxing is pretty insecure on the Mac too.) The fundamental design flaw is that Mac OS X started as an open system, and Apple tried to tack all of these security features on afterward, without otherwise redesigning the operating system. There are two many legacy features, including "must not break" features, too many openings, too much interoperation, too much complexity. Moreover, it often seems to me that Apple security engineers are not particularly experienced with Mac OS X, which is why an old hack like me can relatively easily find things that they've overlooked.
All I'm sayings is it's kinda weird to respond to people who are trying to make a system more secure by saying "well it will never be perfect so fuck it it's just theater". I mean really? I think we agree on everything else.
"Real world attacks (e.g., XCSSET) and researchers have consistently shown that TCC, while often a nuisance to users, does not present a significant obstacle to attackers."
> 21.5% of developers used macOS in 2015,[1] increasing to 27.5% in 2020.
I said the native Mac software ecosystem is slowly dying, and eventually everything on the Mac will just be a cross-platform afterthought. The total number of registered Apple developers (not Mac specifically) isn't really relevant to that. Every iOS developer has to use macOS in order to access Xcode, but that doesn't mean they're making native Mac software.
> We should be making a stink [...] Not just throwing up our hands and saying fuck security I’ll just use Linux.
I'm not sure what you're asking for, here. People have protested OS changes from Apple and Microsoft for decades, but it's never worked. Your only viable method of protest is using an OS where you control the featureset, which (as you've pointed out) is an unrealistic and bad habit.
I wonder why he'd transition to Raspbian and work on a Raspberry Pi.
Maybe to get that nostalgic "let's wait 4 minutes for our 20 line program to compile" feeling again that he must have had in the late 60's and early 70's :P
People: "Linux is not ready for the desktop."
Ken: "You know nothing. Compared to what I'm used to, it's been ready since version 0.01."
Hey now, before the pandemic, you could get as many Pi4s as you wanted. And if you were ok with the 2GB and/or 4GB version, they were usually selling at less than MSRP.
At the Microcenter near me, they kept stacks of them in a cabinet near the cash registers, and would offer them up at a discount when checking out.
That's a lot of devices. I've got six of the Model 4 Rpi and whilst they're fun to play with, they're not ideal in terms of their hardware and it's a pain trying to find a decent USB drive to boot them from to get decent performance (they'd be so much better with an M2 interface). The only one that I use regularly is one that I've got LibreElec installed on and running as a Kodi box.
Before spending time around here, I used to think it would be an embarrassing reveal for most programmers if it came out they used Apple hardware. Like being a WWII expert and rooting for Germany. Couldn't believe that people who could appreciate the beautiful magic of coding and OSS could still even like hyper-commodity computers. But I understand now how naive that was.
"It Just Works" was Apple's value proposition for a long time. Now that's on shakier ground, and other operating systems are catching up on stability to the point where the vastly superior cost:performance of non-Apple hardware increasingly wins out. Linux and Windows OEMs are also catching up on style with options available that don't compromise on repairability.
I was a bit disappointed that most of the questions ignored his talk about a very cool jukebox he built and focused on OS drama.
He built a jukebox with all hit songs he could find in it 1900-2000 and for prerecorded music, got a player piano and sheet music and midi and integrated the whole thing. Touch screens, voice activation and so on. Hardware and software and data hoarding project.
He said he has massive cabinets of CDs, all the music he ripped and tested audio encoders with his own ears.
Ken is 80, and still building cool side projects and scratching his own itch! That's the story.
Be like Ken by building something cool, not by using whatever OS.