These tentacular megacorporations are a problem. Amazon is OK as a merchant, MS as an OS developer, Google as a search engine… If they do vertical integration the market is corrupted.
Vertical integration is when you control the entire product, in consumer electronics Apple is the gold standard; they make the software, hardware, and processors then integrate them into iPhones and macBooks. Tesla is a good example in the automotive space, their goal with the mega-factories is "raw materials in, cars out" and they work to build as many of the parts themselves as possible.
Alternately Microsoft just makes a good enough OS that runs on good enough hardware from commodity vendors, so you get good enough computers. Most auto makers buy good enough components from 2nd and 3rd tier suppliers and integrate them into good enough cars.
That is a great explanation of what vertical integration is. I am not sure I see why it is inherently bad.
I guess a large vertically integrated option could make it hard for alternatives to compete. That is more of a monopoly problem than a vertical integration issue though.
I do agree with interoperability requirements though. I see nothing wrong with Apple offering a fully vertically integrated product. The issue is when I cannot run my own OS on the hardware, my own apps on their OS, or interact with hardware from other vendors.
But that’s exactly the problem. If the company is kind about it, or forced to play nice by effective regulation, there’s no issue. But if there’s no regulation and the company wants to, it tends towards monopolistic tendencies. And there’s nothing that incentivizes a company to play nice forever, in fact they’re incentivized to maximize profit. So Vertical Integration is bad without being checked.
Very dumb complaint, but it’s something I can’t ignore after seeing it: Why are all icons for the sound previews symbolic, but the ones for the notification and USB colored?
You just want to get a feel for it, so I suggest what I've used with success in the past :
Windows host
Virtualbox
Linux Mint with the XFCE desktop environment.
All free, Linux mint is newbies friendly and XFCE is light enough to run well in a VM. It is Ubuntu based so it's very well documented (basically 99% of the tutorials for Ubuntu work with Mint) but it comes with less bloatware and a more ethics.
Of course no single Linux distribution is perfect or we would all be using it but I suggest you don't lose time looking for a distro. Just pick one and install it. If you don't like the look and feel, then try another. You can distro hop through several of them to taste the variations. But the general principles are pretty much the same across the board.
Virtualbox is very easy to use out of the box, even if you have very little experience with virtualization. Everything is in one place and pretty much self explanatory.
Hyper-V is more complicated and requires that you have a Enterprise, Pro or Education license. It cannot be activated on the Windows 10 or 11 home edition.
If we are marking the birth of Linux and trying to call it GNU / Linux, we should remember our history.
Linux was not created with the intention of being part of the GNU project. In this very announcement, it says “not big and professional like GNU”. Taking away the adjectives, the important bit is “not GNU”. Parts of GNU turned out to be “big and professional”. Look at who contributes to GCC and Glibc for example. I would argue that the GNU kernel ( HURD ) is essentially a hobby project though ( not very “professional” ). The rest of GNU never really not that “big” either. My Linux distro offers me something like 80,000 packages and only a few hundred of them are associated with the GNU project.
What I wanted to point out here though is the license. Today, the Linux kernel is distributed via the GPL. This is the Free Software Foundation’s ( FSF ) General Public License—arguably the most important copyleft software license. Linux did not start out GPL though.
In fact, the early goals of the FSF and Linus were not totally aligned.
The FSF started the GNU project to create a POSIX system that provides Richard Stallman’s four freedoms and the GPL was conceived to enforce this. The “free” in FSF stands for freedom. In the early days, GNU was not free as in money as Richard Stallman did not care about that. Richard Stallman made money for the FSF by charging for distribution of GNU on tapes.
While Linus Torvalds as always been a proponent of Open Source, he has not always been a great advocate of “free software” in the FSF sense. The reason that Linus wrote Linux is because MINIX ( and UNIX of course ) cost money. When he says “free” in this announcement, he means money. When he started shipping Linux, he did not use the GPL. Perhaps the most important provision of the original Linux license was that you could NOT charge money for it. So we can see that Linus and RMS ( Richard Stallman ) had different goals.
In the early days, a “working” Linux system was certainly Linux + GNU ( see my reply elsewhere ). As there was no other “free” ( legally unencumbered ) UNIX-a-like, Linux became popular quickly. People started handing out Linux CDs at conferences and in universities ( this was pre-WWW remember ). The Linux license meant that you could not charge for these though and, back then, distributing CDs was not cheap. So being an enthusiastic Linux promoter was a financial commitment ( the opposite of “free” ).
People complained to Linus about this. Imposing financial hardship was the opposite of what he was trying to do. So, to resolve the situation, Linus switched the Linux kernel license to GPL.
The Linux kernel uses a modified GPL though. It is one that makes it more “open” ( as in Open Source ) but less “free” ( as in RMS / FSF ).
Switching to the GPL was certainly a great move for Linux. It exploded in popularity. When the web become a thing in the mid-90’s, Linux grew like wild fire and it dragged parts of the GNU project into the limelight wit it.
As a footnote, when Linus sent this announcement that he was working on Linux, BSD was already a thing. BSD was popular in academia and a version for the 386 ( the hardware Linus had ) had just been created. As BSD was more mature and more advanced, arguably it should have been BSD and not Linux that took over the world. BSD was free both in terms or money and freedom. It used the BSD license of course which is either more or less free than the GPL depending on which freedoms you value. Sadly, AT&T sued Berkeley ( the B in BSD ) to stop the “free”‘ distribution of BSD. Linux emerged as an alternative to BSD right at the moment that BSD was seen as legally risky. Soon, Linux was reaching audiences that had never heard of BSD. By the time the BSD lawsuit was settled, Linux was well on its way and had the momentum. BSD is still with us ( most purely as FreeBSD ) but it never caught up in terms of community size and / or commercial involvement.
If not for that AT&T lawsuit, there may have never been a Linux as we know it now and GNU would probably be much less popular as well.
Ironically, at the time that Linus wrote this announcement, BSD required GCC as well. Modern FreeBSD uses Clang / LLVM instead but this did not come around until many, many years later. The GNU project deserves its place in history and not just on Linux.
Something is open source or isn’t. There’s a set, binary definition.
I get the feeling you’re implying a difference/aversion between those two terms which doesn’t exist. This and the combination with a nonsensical statement about amount of GNU packages vs non-GNU packed makes it feel like you’re pushing an agenda here: There’s far more free software than just GNU’s - that’s a success for free software and the GNU project. There’s no connect between the argument you’re obviously implying.
Also HURD never took off - but why should it? The GNU project’s goal is a fully free operating system, with Linux being persuaded to adopt a proper license there’s no real need for HURD. It doesn’t mean it isn’t a fun project.
Which two terms? Everyone has an agenda but I am not sure what I am being accused of here. Do you mean Free Software vs Open Source? The FSF goes to great lengths to distinguish between those two terms:
The BSD license allows incorporation of BSD code in non-free projects. That was both an advantage for capitalists while simultaneously moving hobbyists away from it’s development. Kind of an important bit of info.
First start using ardour gimp inkscape libreoffice and blender on windows, then dual boot or use a VM to install Linux and start challenging yourself to use it for real stuff when you can. You may eventually realize you’re using Linux much more than Windows, like when you boot into Windows and every time require a bunch of updates. Eventually your windows will be so out of date you remove it entirely or start using a windows vm.
Yes, but those minor traces are easy enough to remove, especially if you don't care about being "ceritified" by Google (i.e. are not planning to run the Google services).
If my device is compatible, does it automatically have access to Google Play and branding?
No. Access isn’t automatic. Google Play is a service operated by Google. Achieving compatibility is a prerequisite for obtaining access to the Google Play software and branding. After a device is qualified as an Android-compatible device, the device manufacturer should complete the contact form included in licensing Google Mobile Services to seek access to Google Play. We’ll be in contact if we can help you.
Google services are entirely missing from Android open source. The Google Play package is what contains the entirety of Google’s services.
Not sure if anyone remembers but back when cyanogenMod was the go-to, early versions had Google services included. Google sent a cease and desist notice and said it was a license violation. You cannot distribute it as part of the OS by default. The next release of cyanogenMod had it removed. Users had to flash the package if they wanted it.
Right but the topic was about google’s data harvesting and what I meant was that you can’t just grab any AOSP distribution if you want to minimize that, you need to pick one that replaces the parts that send data to google. LineageOS for example still phones google for quite a number of services.
As far as “easy to remove” goes, I think that’s kind of debatable if you want to do it in a way that’s sustainable long term considering the effort that goes into e.g. GrapheneOS or DivestOS.
Edit: here is a list of the kind of stuff you need to watch out for if you want to minimize the data sent to google
I was answering under the assumption/the context of of "Amazon wants to release an Android-based OS that doesn't contact any of Googles services".
So, when I said "easy enough to remove" that was relative to releasing any commercial OS based on AOSP, as in: this will be one of the smallest tasks involved in this whole venture.
They will need an (at least semi-automated) way to keep up with changes from upstream and still apply their own code-changes on top of that anyway and once that is set up, a small set of 10-ish 3-line patches is not a lot of effort. For an individual getting started and trying to keep that all up to do date individually it's a bit more of an effort, granted.
The list you linked is very interesting, but I suspect that much of that isn't in AOSP, my suspicion is that at most the things up to and excluding the Updater even exist in AOSP.
Am I understanding correctly that you are building the image by copying in key elements from the host machine’s functioning nginx installation?
This is creative but not common approach to docker.
Normally software is installed following the officially documented procedure (imagine installing using apt or a shell script via RUN). Sometimes software documentation has specific recommendations to follow for containerized installs.
It’s common to have the version defined as a variable where a change in value invalidates the docker layer cache. To me it’s unclear how caching would work with your dockerfile, for example, in the event of a upgrade. You could also see how a breaking change (such as one in the paths you are copying) could run into issues with your hardcoded approach.
In the case of software like nginx, I would use the official image, mount config/cert files instead of copying, and extend in my own dockerfile if needed.
Not from the host machine, but from the official nginx image ( nginx:mainline-alpine3.18-slim ). And what it (basically) does is separate the essential commands/files inside a scratch image and gives every command a custom username tag.
A bit late but you might want to have a look at docker multi-stage build documentation which does exactly what you did (start from a base image then copying stuff from it to your own image), something like that:
Which will simplify building new images against newer “build” image newer tags easier.
btw, you were quite creative on this one! You also might want to have a look at the distroless image, the goal being to only have the bare minimum to run your application in the image: your executable and its runtime dependencies.
You’re welcome! scratch and distroless are indeed basically the same thing, scratch being the ‘official’ docker minimal image while distroless is from google - as I’m more a Kubernetes user (at home and at work) than a Docker user, I tend to think about distroless first :) - my apologies if my comment was a bit confusing on this matter.
By the way, have fun experimenting with docker (or podman), it’s interesting, widely used both in selfhosting and professional environments, and it’s a great learning experience - and a good way to pass time during these long winter evenings :)
One that is relatively up to date with their graphics drivers. Then just install steam/lutris flatpaks and go crazy. Performance difference is pretty much negligible once it’s set up.
for 1, in linux no output is often indicitive of no problem. To verify if your previous command exited successfully, type ‘echo $?’ at the command line and if its anything but 0 its an error.
For 3, I do the same but since I’m the only user I auto login so its still just one password to enter to get to a desktop.
Depends if you’re using a graphical login manager or not. If so, you’ll have to search the name of it and ‘autologin’ in your favourite search engine. Its typically no more then checking a box and adding your username.
I dont use a graphical login manager, I just let it boot up and agetty (from util-linux) logs me directly into my shell (because I added -u ’ to the config.). Then my shell profile takes care of starting the graphical environment for me.
Its just personal choice but I dont see any point in a login manager when Im the only one logging in. I understand that it may come as part of the desktop suite though. I prefer to start with nothing and add what I want versus getting everything and removing what I dont want
linux
Active
This magazine is from a federated server and may be incomplete. Browse more on the original instance.