So, OS’s that don’t ship with GNU components aren’t Unix?
See: any BSD.
All of the things you mention in your post have _alternatives_ that also comply with whatever they reference: there are other MP3 encoders, there are other types of gas, there are other democratic republics…
That's so true. The OS ships with certain things preinstalled, but you can install other or newer versions of them.
And if I found myself in a shop where I needed to use standard command line tools to do my job but wasn't allowed to install or update them, I'd be looking elsewhere because life's too short for that bureaucratic overreach.
I hear horror stories like bosses not allowing their employees to choose their own tooling. Pretty much any office like that which doesn't have armed soldiers guarding the building deserves ridicule.
It is, after all, a great tradition of UNIX™ systems to install the OS... and then immediately install GNU tools on top to make it less painful to work on. This is just macOS carrying the banner proudly:P
The linked page lists some SCO, IBM, and HP products. I have no idea how relevant any of those are. Is macOS the only remaining Unix with significant a significant user base?
It's absolutely wild. They've broken multicast and UDP for all applications that do not pay Apple money (like open source, community coded ones). Which means things like mouse/keyboard sharing applications like synergy/barrier/etc stopped working in many cases. But that's just the tip of the iceberg. And there's no way for a human to control this "multicast entitlement" on their own OS how they like it. The only way for applications on Sequoia to have multicast network access is for some commercial developer pay for it.
This is very not Unix. Careful out there "up"grading.
Both of your links refer to that entitlement being required on iOS, not macOS, are you sure it applies to Mac (and if it does, does that even matter given non-app-store-apps can assert whatever entitlements they'd like with impunity)?
Your link is about iOS. macOS is not concerned. Also, that makes sense to have a UNIX certification for macOS, not much for iOS, which is a different kind of platform.
It sounds pretty classically Unix to me. If the entitlement comes on a slip of paper they have to mail you, perhaps in an otherwise empty box or a FedEx envelope, it REALLY sounds like Unix.
Is it possible to get around the authorization requirement if you don't get your app notarized? Obviously that makes it harder to distribute it and to convince people to install it.
The slow crawl of locking down Macs has made me consider abandoning the platform. I think I’m done with Macs. Might hold on to iPhones, but I want to be able to use my phone however I want, too.
Sincere question: why would those kinda of apps need multicast? I don't know much/anything about them but it seems to my naive eye like they'd use plain old unicast UDP.
Those who choose to buy, use and develop for Apple systems supposedly know what they want and deserve to get it good and hard - free after H.L. Mencken
I remember a time when this mattered in terms of branding (Mac OS X being a Unix). It meant that the OS was “industrial-class”. This was against a backdrop of Macs not being seen as serious machines, unlike Windows machines which were more “corporate”.
But nowadays macOS is so pervasive that a Unix certification doesn’t buy much brand cachet. Perhaps it might help for specific compliance requirements but those are likely the minority.
> It meant that the OS was “industrial-class”. This was against a backdrop of Macs not being seen as serious machines, unlike Windows machines which were more “corporate”.
I love how Apple basically cp -R BSD, became the most profitable company in the world off it and gave nothing back except a walled garden and lock in. But hey it’s still technically Unix.
To be fair, all of those started as FOSS outside of Apple. LLVM came from academics at UofI, CUPS has been around since forever and they just bought it and started paying the main dev, Darwin came from pillaging NeXT, and Webkit famously was based on KHTML from KDE.
It still drives me up the wall when people think that Apple created Webkit. Just another FOSS project that Apple forked. Not only that but they didn't want to play nice with the original project so KHTML was dead within years.
I never said they created it. I said they have made substantial oss contributions.
And does it really matter? If an OSS project is forked, soars to new heights, and sees use in millions of devices all over the world is that a bad thing? Or would you prefer it remain a quiet little sleepy project for 3 devs to bikeshed on.
Webkit was not a simple fork. Apple took KHTML, pretended to upstream patches for an extremely short amount of time, then purposely made it too difficult for the original maintainers to accept patches any longer until they had successfully killed the project. Classic take what you need from other's hard work at all costs.
It's a lot like how whenever Apple adds a new feature to one of their operating systems, for Apple and its most dedicated fans, that feature is Apple's invention.
Much like macOS' very official and increasingly nominal affiliation with Unix, Apple's involvement with open-source of more interest to people who see that as a feather in Apple's cap than to people who see open-source as a first-order value of their own. Apple has little concern for either user freedom à la the free software movement or for the collaborative modes of development touted as the primary benefits of open-source by the open-source movement, and this shows in their historical and present actions. For people who give a shit about the goals, practices, and principles of open-source or free software, the creation of WebKit is remembered and remembered as a fiasco.
People who Apple fans first don't know or care about things like 'upstream-first' development, what characterizes a good citizen vs. a bad one in the open-source world, or likely even that companies other than Apple are also criticized by the F/OSS community for engaging in similar behaviors (e.g., the scandal provoked by Canonical's release of Mir after it was developed for a prolonged period in secret).
If you actually look at Apple's history with these, it's embarrassing for Apple.
Apple effectively abandoned CUPS when Michael R. Sweet left the company, which is why the OpenPrinting fork exists. For something like a year they let the project languish and accepted virtually no patches, either. No one but Apple uses Apple CUPS anymore.
Darwin is subject to periodic, incomplete code dumps to the point that for decades even getting it to build has been a struggle that includes tons of downstream work. At the moment, such efforts are in a historic slump; no recent stack is usable for anyone and it's been that way for years.
WebKit was a hostile fork (without any history of antagonism to motivate it) of KHTML, essentially classic code-over-the-wall stuff. Its original announcement was cause for consternation in the F/OSS community, not celebration.
LLVM I don't know much about the governance and maintenance of. Maybe someone else can speak to that.
But as useful as LLVM and WebKit have been to downstream projects, Apple's record with open-source is mixed at best.
Ah CUPS, that software which just received some CVEs for decades old security vulnerabilities
Also, Darwin is more of a hypothetical thing. There was a time when they actually released CD images IIRC, but that was a long time ago. I haven’t seen a pure Darwin installation in probably almost 20 years now.
I’ve always thought POSIX compliance is the primary thing vendors should be targeting for portability. I was trying to find something that would explain the difference between POSIX and UNIX compliance, but there’s not great sources on this specifically.
I imagine UNIX is a superset of POSIX, I’d be curious if anyone knows of a simple comparison between the two?
I’ve always wondered, since back when OS X 10.5 got apples first Unix certification.
What does that help you as a developer? Can you just download the spec for free, built against it and have a working app for several platforms that not only compiles and runs but also behaves the same on all of them?
55 comments
[ 2.7 ms ] story [ 18.8 ms ] threadSee also: Lame ain't an MP3 encoder, natural gas, DPRK
See: any BSD.
All of the things you mention in your post have _alternatives_ that also comply with whatever they reference: there are other MP3 encoders, there are other types of gas, there are other democratic republics…
And more importantly, GNU isn’t Unix.
Why would you write this? Are you trolling, or do you believe that's really what was meant?
https://en.wikipedia.org/wiki/Denying_the_antecedent
And if I found myself in a shop where I needed to use standard command line tools to do my job but wasn't allowed to install or update them, I'd be looking elsewhere because life's too short for that bureaucratic overreach.
I hear horror stories like bosses not allowing their employees to choose their own tooling. Pretty much any office like that which doesn't have armed soldiers guarding the building deserves ridicule.
HP-UX is still being released, not sure who is using it but it must bring in enough revenue to justify support.
SCO and whoever they sold to have been dead for quite a while.
It's absolutely wild. They've broken multicast and UDP for all applications that do not pay Apple money (like open source, community coded ones). Which means things like mouse/keyboard sharing applications like synergy/barrier/etc stopped working in many cases. But that's just the tip of the iceberg. And there's no way for a human to control this "multicast entitlement" on their own OS how they like it. The only way for applications on Sequoia to have multicast network access is for some commercial developer pay for it.
This is very not Unix. Careful out there "up"grading.
Although that might imply there must be a business involved that is wronged by their actions... and many open source programs do not have that luxury.
"Apple is bringing some networking protections over from iOS to macOS starting with Sequoia. Please see the following links: https://developer.apple.com/videos/play/wwdc2024/10123?time=... https://developer.apple.com/forums/thread/663858 https://developer.apple.com/videos/play/wwdc2020/10110/ "
I'm already fielding questions about this from affected Mac OS Sequoia users in the #barrier IRC help channel.
You can build local open src software without this.
But nowadays macOS is so pervasive that a Unix certification doesn’t buy much brand cachet. Perhaps it might help for specific compliance requirements but those are likely the minority.
See FoxTrot comic from February 2002:
* https://www.gocomics.com/foxtrot/2002/02/25
- LLVM - https://llvm.org/
- They are the stewards of CUPS (Common UNIX Printing System) - https://www.cups.org/
- Darwin kernel is open source - https://github.com/apple-oss-distributions/xnu
- Webkit - https://opensource.apple.com/projects/webkit/
That said, the other contributions like WebKit and LLVM have had a wide and positive impact.
And does it really matter? If an OSS project is forked, soars to new heights, and sees use in millions of devices all over the world is that a bad thing? Or would you prefer it remain a quiet little sleepy project for 3 devs to bikeshed on.
Much like macOS' very official and increasingly nominal affiliation with Unix, Apple's involvement with open-source of more interest to people who see that as a feather in Apple's cap than to people who see open-source as a first-order value of their own. Apple has little concern for either user freedom à la the free software movement or for the collaborative modes of development touted as the primary benefits of open-source by the open-source movement, and this shows in their historical and present actions. For people who give a shit about the goals, practices, and principles of open-source or free software, the creation of WebKit is remembered and remembered as a fiasco.
People who Apple fans first don't know or care about things like 'upstream-first' development, what characterizes a good citizen vs. a bad one in the open-source world, or likely even that companies other than Apple are also criticized by the F/OSS community for engaging in similar behaviors (e.g., the scandal provoked by Canonical's release of Mir after it was developed for a prolonged period in secret).
Apple effectively abandoned CUPS when Michael R. Sweet left the company, which is why the OpenPrinting fork exists. For something like a year they let the project languish and accepted virtually no patches, either. No one but Apple uses Apple CUPS anymore.
Darwin is subject to periodic, incomplete code dumps to the point that for decades even getting it to build has been a struggle that includes tons of downstream work. At the moment, such efforts are in a historic slump; no recent stack is usable for anyone and it's been that way for years.
WebKit was a hostile fork (without any history of antagonism to motivate it) of KHTML, essentially classic code-over-the-wall stuff. Its original announcement was cause for consternation in the F/OSS community, not celebration.
LLVM I don't know much about the governance and maintenance of. Maybe someone else can speak to that.
But as useful as LLVM and WebKit have been to downstream projects, Apple's record with open-source is mixed at best.
Also, Darwin is more of a hypothetical thing. There was a time when they actually released CD images IIRC, but that was a long time ago. I haven’t seen a pure Darwin installation in probably almost 20 years now.
I imagine UNIX is a superset of POSIX, I’d be curious if anyone knows of a simple comparison between the two?
What does that help you as a developer? Can you just download the spec for free, built against it and have a working app for several platforms that not only compiles and runs but also behaves the same on all of them?