Linux Has Not Won, Microsoft is as Dangerous as Ever, Fie on Secure Boot
lxer.com
lxer.com
In the last 20 years, Linux has taken over many domains. First servers. The big win that made Linux impossible to ignore. Ten years later a major shift to the userland was mobile: there are half a billion Android devices sold. Not to mention embedded systems. Meanwhile, Linux has also become the way to build supercomputers: the world's fastest supercomputers these days mean assembling huge Tesla GPU racks to run Linux.
While I do realize that secure boot can make Linux's life harder, I also realize that Linux per se isn't going anywhere or isn't attacked by Microsoft but general purpose computers are suffering from exhausted momentum. There are more and more devices that come with a preinstalled system, such as Android and iOS and the traditional desktop/laptop paradigm of buying hardware and installing whatever you want to run on it will continue to move into the marginal.
I predict that eventually it's the programmers who will be the only ones who exercise tasks of so varying nature that they need a more complex interface, something like what I'm using now. Most of the rest of people can manage with mostly touching a screen to use applications that offer a set of predefined capabilities, occasionally hooking up a keyboard.
The era of immutable operating systems with applications installable from a prefiltered app store is probably what actually works for most people. It's us programmers who long for the days when hardware was to be bought and software was to be written because we know that the lowest layer must always be there. For us, it was the very definition of a computer: programming hardware was all there ever was to computers. But most people don't need that.
Most people are perfectly happy with an appliance. They don't want to install updates or manage their system, they just click the power button and immediately continue from where they left the last time. They want an application and they tap it and it gets installed without clicking through a dozen pages of a helper wizard or aptitude install commands. And appliances such as tablets and phones and touchscreen laptops deliver a much better experience for non-programmers.
What we as programmers see as a bad thing is that programming will be removed farther from the end users. We all know how we started with playing games and ended up writing our own games because our computers allowed, and to an extent, suggested that. We would want to preserve that heritage to the future whizkids and future programmers. But that's still got a perspective bias: most people so far still haven't "found it" even if their computer would have allowed it. It's just us for which "finding it" was the revelation, and we know we would've "found it" anyway, somehow.
Maybe they'll be selling a limited stock of programmer's computers in ten years that are fully programmable. There's definitely a market for those because you can't write the nice appliances without a real programmers' computer. Maybe my grandkids will receive one of my old programmer's workstations from my work place, and use that deprecated hardware to teach themselves programming and find it a thousand times more interesting than playing games on their phones, much like my predecessors managed to hook themselves up with cheap nighttime computing time in a mainframe at their father's workplace and abused those cycles to write games to entertain themselves.
It strikes me that with better virtualisation technology you could have your "appliance" device and your dev box sitting side by side on the same piece of hardware (or access it remotely). The question will be more about access to that, will it be something you can just install like everything else on your device or will you need to do some complex jailbreak or sign an agreement with the manufacturer to install?
I'm not sure to what extent Linux has "take over servers" though. Surely this would seem to be the case with HTTP servers (which are the most visible to the internet) but if you go into businesses from small to large you find an awful lot of Windows Servers doing stuff like Active Directory, Exchange and File/Print sharing. In fact almost all of the "IT Service Provider" companies in my area are Windows only shops.
15-20 years ago for a lot of functions a big UNIX box would be installed. More often than not, that box is now a Linux box. Linux didn't so much as push Microsoft out of the datacenter as replace UNIX.
There were recent discussions about AD here in HN.
SQL Server is a great piece of software, better than Oracle and MySQL and its only fault is to run on Windows.
We need to get rid of all the kinks out of linux, make plug and play work seamlessly, and improve usability. Part of that means getting rid of the dependency to use the terminal (who uses that piece of shit anyway?) and the need to config text files to fix driver issues. Linux still has the ability to reach #1 in non-programmer desktop usage.
By combining both of your comments, I do think the future is clear: the only truly free, general purpose computing machines will run Linux, and will _look_ like an appliance, just because of all the locked down walled in appliances Linux will be competing with.
It won't hurt that this will make Linux easier to use. And Gnome 3 is an example of how not to make it easier to use – give the user what they want, do not take it away.
Actually, it's not. You just disagree with it. Sometimes when you disagree with someone's conclusions, it's hard to see the merit of their reasoning.
We need to...
There are two complications to this. One is that the "we" of Linux development has never and will never have the kind of coherence that it takes to appeal to an extremely large audience. (Even calling it "Linux" isn't always agreed-upon). There are a great many people working on Linux or making things from Linux, but they are not all going in the same direction. This is both a strength and a weakness.
The second is not so much a different reason as it is an example of a major weakness of this style of development: It's a demonstrably inefficient way of addressing the needs of certain kinds of users. The list of things you say need to be fixed with the Linux desktop, which is quite typical and also quite incomplete, has also remained constant for more than a decade as year after year "of the Linux Desktop" passed by. Although the situation has improved, "We" have not moved on to a better class of problems because most people working on them want to use the terminal, or at least don't mind editing text files for configuration, and correspondingly are not very good at addressing the needs of people very much unlike them.
Okay, I joke, I know several people who run linux on their desktop.
It seems like something that you grow out of, though. I use a mac, and most of the people I know who /really/ know [1]linux do too. In fact, the person I know who "knows linux" the most inside and out runs an ancient PPC macbook. I think he said he installed Ubuntu on his machine at work, though.
The reason for this is that we all run linux* on our servers. My primary machine is a VPS from linode which runs gentoo linux, and has for a very long time. The aforementioned PPC macbook guy doesn't really have a primary machine anymore. He is getting into power electronics, and is more likely to be sitting at an oscilloscope than a laptop.
Lately, my hobby box has been a raspberry pi which I'm looking to make it into a telepresence robot for our lab. Interestingly, live streaming video from a linux machine doesn't quite seem like a "solved" problem yet. It's lots of twiddling ffmpeg to make things work; things still aren't "plug and play".
This is cool, by the way, it feels like the first time I stayed up all night installing gentoo.
--
I guess that for my peer group, linux was something that felt really revolutionary and new and important 10 years ago. We were battling against the big bad microsoft, THIS was going to be the year of the linux desktop! (No, MOM! You have to install Suse! It's totally easy let me show you!).
Then...we won.
Linux is just a tool for me. Like a hammer. When I want to use it, I pick it up and use it. It doesn't feel revolutionary and important anymore because it just feels /obvious/.
[1]: It's a bit silly to say "linux". I'm generally not specifically using any of the features of the linux kernel. Most of what I'm doing could easily be replicated on a *BSD system. Although this whole "gnu plus linux" debate thing is a bit silly. When I say "linux" people know what I mean. Sorry, Stallman :(
On the contrary to how you described it - when people grow out of Mac OSX and Windows, they start using Linux on the desktop.
And of course if you are at all involved in tech you probably don't have a really representative sample.
See http://en.wikipedia.org/wiki/Linux_adoption#Measuring_deskto...
I can't think of a strong reason to suggest that it would be higher. The argument is made on that page is that it is strong in places like China, though my understanding is that is not correct. Plenty of PCs in china may sell with Linux (or freeDOS) but my understanding is that they are always ubiquitously flattened and replaced with pirate Windows.
My brother, who I don't consider a tech person, switched over to linux after seeing me run it for years and needing a faster OS for an old laptop. He now uses it even on his newer machines. Along with my brother I have had a few other family members ask about maybe switching their computers to linux.
With more devices based on linux such as ultrabooks (chrome os) it looks like more companies will be writing drivers for linux; one of the larger issues with linux. With Valve moving to linux we have seen huge improvements in graphics card drivers which is awesome, and hopefully more game companies will start working on linux ports.
The big question though, is whether or not linux moving to consumer type interfaces is a good thing. See gnome 3, unity, etc. making things easier to use on the surface, but at the same time removing lots of the configurability and power that many linux users enjoy.
Here is a good review of shifts in ideas behind various Linux DEs:
http://www.datamation.com/open-source/experimentation-vs.-tr...
I agree with the author that KDE made the wisest choice not to narrow options and to embrace various form factors with different design approaches (instead of lumping it all together into one mobile style UI).
I know that at the low end it's sometimes installed on computers which are given to very tech unsavvy people who don't need anything outside of a browser and can't seem to touch a Windows PC without acquiring 300 viruses.
I have almost zero familiarity with Linux and dread the prospect of having to learn a new OS on my everyday computer, but Windows 8 is just THAT repulsive.
If he were around to see this, he'd consider it a stolen business model and start lawyering up.
If you are into games and are looking to run Steam, I suppose you want more horsepower. I usually buy stuff off of Newegg and build my own if it's not a laptop.
I would not say that there is a lot to learn especially except for some of the plumbing (networks for instance) that works differently.
The walled garden is not new. If you need Office for even 1 thing, have to do work on just 1 IE only website, can't live without even 1 DirectX powered game or do any MS specific dev (.NET, SQL Server, etc) you're already stuck on Windows for at least 1 machine. If switching between OSes is too painful, as it is for most, then you're stuck on all machines.
Dependents inherit all dependencies. Dependencies are like a helpful drug with bad side effects, only what is necessary should be accepted. And like with drugs, there are distributors, consumers, pushers and addicts. Which labels apply just depends on everyone's intentions.
This just seems silly. How is this an "attack", does the Linux foundation somehow lose funding proportional to the number of Windows licenses sold?
Who said Linux has won the desktop?
So now it's got a bit harder to remove the Windows 8 license you just voluntarily paid for. Oh dear.
If you are going to whine about Secure Boot please write an article based on its technical merits.
>> I guess, under 10,000 hours of testing
I would of thought that the regression of Microsoft is inferior to the diversity of people using Linux daily. Another benefit to a rolling release cycle I suppose
It's not voluntary as long as the option of not buying doesn't exist. It's coercion.
See Ubuntu Bug #1 https://bugs.launchpad.net/ubuntu/+bug/1 - Microsoft has a majority market share
So if you believe the graph that someone else posted today, Microsoft has already lost and that means Ubuntu/Linux has won... or something.
http://www.businessinsider.com/mary-meeker-2012-internet-tre...
From the article:
>Every purchase of a Windows license is an attack on Linux. Linux has not won, and Microsoft is as dangerous as ever
What if people buy an iPad instead of a Windows PC? Is that an "attack" on Linux? How about total lack of developer freedom in iOS or even charging for in app purchases or banning linking to a website for buying things?
Or even if the user buys a Mac instead, do you think Apple would be better at enhancing user freedom that MS was? How about the Apple tax on all hardware that you need to buy to get Mac compatibility if it gets a monopoly?
Now I have a concern again, and that doesn't feel nice.
You are not the desktop market, just a single user.
Oh, Linux community, shine on, you crazy diamond!
For what it's worth, I think Microsoft are doing this for the security reasons they give. But I think they were well aware of the hurdle it would present to Linux distributions when they chose to do it.
And nobody will have noticed.
I doubt an Android tablet will __ever__ be able to process my post-production renders in a reasonable amount of time.
This is the #1 thing I do with my desktop computer at home.
A lot of the video work was done on co-processor boards (Media 100 and Avid), but almost all of the After Effects renders were done using the base Mac.
My first job was to supervise a render of an animated banner on a commercial, it took 48 hours to render 30 seconds worth of video. I had to be there the whole time to ensure that a crash didn't stop the render.
Things change and it is foolish to underestimate what CPUs and GPUs will be capable of. There are some pretty stunning games taking advantage of the iPad's GPUs, I would surprised if tablets in 5 years couldn't do what your professional workstations do today (the problem will be that professional standards will go up in the meantime and you'd turn your nose up at what you can do today).
No need to to have that be your primary machine. Or to burden your render farm with the task your table will be better suited for.
Sure! Would you be willing to provide the necessary funds?
I think this that is just wrong, can someone please verify? I thought the user could add their own master keys to boot their own OSes apart from being able to disable secure boot as mandated by Microsoft for Windows certification (as much as it can force OEMs without the anti-trust rulings stopping it from mandating requirements to OEMs).
There is so much FUD and misinformation spread by folks who you would think be otherwise smart and knowledgeable that it's hard to find to what to believe and what not to.