Arguing the future of Linux.. and death of the PC...

 I found myself recently tangled up in a discussion on Facebook about Linux and a suggestion that perhaps the only way forward for it as a desktop operating system is with something like ChromeOS.
I suggested, fairly bluntly, that I'd go back to using FreeBSD on my desktop before I'd put up with ChromeOS.

The poster came across as very much a fanboy of Linus Torvalds' - the creator of the Linux kernel - a guy known for his outbursts at contributors and other trantrum-esque behaviours. A guy to whom many are indebted, yet we simply don't always agree with him.

The poster was insistent that Linus has a point, because GNU/Linux has become so fragmented with so many distributions and so many different options that it's simply too much work for vendors to deal with Linux.

I pointed out that vendors that *do* deal with Linux, typically offer a pair of packages - usually the same piece of code packaged via one of two different methods that facilitates a nice, simple installation on the bulk of Linux distributions in use. He still insists on arguing the point.

I further point out that there are plenty of vendors who do provide applications for Linux - entire office suites, graphics editing and so on - these are cross platform products that work on Windows as well, so there are vendors out there. And these ones make their applications open source... as opposed to the likes of Adobe, who will charge as much as they can for a product.

The issue of fragmentation within the Linux community has nothing to do with the likes of Adobe supporting Linux - Adobe make commercial products sold at premium prices. The nature of Linux users isn't in paying significant amounts of money for software, so why would they? It's full of technically oriented types who they see as less inclined to pay (if we wouldn't pay for an operating system, why would we pay for their software - it's the exact opposite of what they think of Mac users). If we need support, we'll drive them nuts - we're typically going to give their Tier 1 and 2 script readers hell and their devs generally don't want to be the ones stuck interfacing customers - especially when many of them would be telling them how woeful their applications actually are.

For the smaller vendors it's somewhat different. There are a lot of applications that have been written in higher level languages that have prerequisites that make them predominantly suited to Windows. Given that Microsoft in particular do make some of those "prerequisites" available to Mac as well, the vendors may find it easy enough to port their application across to Macs. Ultimately though, their choice of programming language and how they've written their application may be the limiting factor. It might be a simple case of - it's just the programmers chosen language. It might be that they chose that language because it provides libraries that provide extensive functionality required for their application. What's happening with Linux has nothing to do with it. It's just a reality of programming.  When I was in high school, programming was taught using Microsoft's Visual Basic and QBasic before that. I lost interest in that about the second I started using Linux.

At the end of the day, a lot of Linus' fanboys either have short memories or haven't been around in Linux land that long. Personally, I've been able to use Linux as my desktop OS exclusively since 2009.  Yes - I do have Windows as well - but I'll explain that in a minute.
The availability of suitable application alternatives these days is impressive compared to what it was when I first installed Linux in 1998. It's still impressive compared to the experience when I tried to use it as my desktop OS in 1999, 2000, 2002 and 2004. Back in those days, a lot of those packages required you compiling from source code. That could take a lot of time, and often become tedious as you worked your way through resolving dependency issues manually. Anything "big" was a pain in the rear. Don't get me started on drivers - especially at the end of the 90s when drivers started taking on the work of hardware to bring down component costs.

It's by no means "perfect" - some of the available applications aren't as good as the Windows equivalents. Yet at the same time, I also didn't have to pay for them so I'm not complaining. You don't expect Bentley quality when you've paid for a Hyundai. You just expect it to get from Point A to Point B.

I mentioned Windows. As a system administrator, part of my reality is deciding on "the best tool for the job". I'm professionally a Linux administrator - I don't do much with Windows. It's not my forte. If a use case is best suited to Windows, then that's what it gets. If it's best suited to Linux, then that's the direction we start with. That then gets nailed down further about which distribution. I spend a lot of time working on RHEL - it has its strengths. Especially with systems where we need things to remain fairly consistent and stable for a long time. If we need something more dynamic, it'll likely be Ubuntu. It's stable, really, really well supported, and Canonical provide support options which makes it a good choice in an enterprise regardless of what the Linux purists - many of whom have never had to walk into a Data Centre - would like to think.

I run Windows at home. Mostly in virtual machines. I keep a virtual machine that I use for remote work. Ironically at work I use a Linux machine even though it's a Windows-centric network, but it's simply because of how their choice of VPN behaves. If I could be bothered manually doing a heap of scripting, I could get around it. I can't be bothered. I don't have the time nor the interest in adding layers of abstraction to what is essentially just a tool that allows me to move from one place to another to achieve an objective. If anything gets even remotely suspect with it, that VM gets nuked and rebuilt. Simplicity is key. I also have a couple of Windows servers running. One because I use Veeam for backup. I also use it at work, so there is some benefit in using it at home. The other, because I have a test domain controller. None of it "necessary" from a personal computing view point. Only "useful" from a professional perspective.

I have a couple of Windows machines I use for my radios. They came with Windows, and to be honest, from a radio perspective, they're not a lot of work. Install program, run and it works. There are Linux options to achieve the same result, but then I'd have to do a heap of additional tinkering to achieve a couple of my objectives. Then I'd have to do a heap of script reworking. Simply, that's starting to make having radios connected to computers feel like just another day at work. 20 years ago, it wouldn't have been a question. If they could run Linux, they'd have been running Linux. These days I simply don't care. I'm far more interested in sending some SSTV pics than I am worried about what operating system I'm running. If Windows breaks on one of them, I'll probably consider moving them to Linux.. especially once Windows 10 support circles the drain. Until then, it's not broken, so I'm not fixing it. I simply don't care *that much* any more.

I learned a long time ago to give up on the delusion of trying to push Linux on to everyone, or trying to justify it. It's a lot of misguided effort for little return. For many - including many IT "professionals", it's all too hard, too unfamiliar or whatever, so they're just going to keep doing what they've always done. For the folks who run applications online, provide infrastructure and so on, it doesn't need justification. It's just a matter of whether or not its the best fit for a specific use case. People with the right inclinations will discover it in their own time.

For those of us who are absolutely comfortable in using Linux as our "daily driver" - it's all of those choices we have available to us - different Window managers, different distributions, kernel options and so on and so forth that put us in that situation. The very fact that we have all of those options is the point - people put in some seriously hard work to make them because they wanted something that suited themselves compared to what was already there. We're typically Linux users because we're not interested in being restricted in choice like we are with Windows or with Macs. What would limiting us to ChromeOS do? It might become a commercial competitor to Windows, but to what end? Sure, it runs on cheaper hardware that really isn't up to running anything else. Right now it's the only reason that anyone uses it. It's good for running a web browser.. not much else.  The hardware it's sold on though certainly isn't up to the task of say Adobe Premiere, even if it run on Linux.  Chances are, if you have applications that need horse power and can justify the high price tag of the software, you're probably going to be willing to invest equally in the box under it. That's why most content creators are seen using Macs rather than sub $1000 HP laptops running Windows.

ChromeOS has its place in the market and might even help pump up the numbers of "Linux users", but there's no motivation for the entire industry to invest into what are the bottom feeders of the personal computing world. And no manufacturer in their right mind is going to put a niche OS like that on serious hardware - it'll come out with Windows because that's what most people want, and the Linux fanbois and girls are likely to just do their own thing regardless - and chances are, it won't involve ChromeOS either.

The internal infighting in the Linux community serves little purpose beyond displaying passion. Those of us that have been here for a while though can generally appreciate how far we've come. Personal computing was a growth industry back then. Now younger generations are less interested in having a screen and keyboard, replacing it instead with a phone screen.. or a TV and a game pad. We've been the alternative for a long time. Linux is embedded in so many devices now that the numbers are unfathomable. It's achieved greatness. I'm not even sure what the value is in trying to streamline it back into a more limited operating system these days. By the time we get there, desktop computing will be dead outside of business. Business will use whatever the bulk of their staff and vendors can support and use (this is how Windows has done so well).

As more applications become cloud based, interfaced through web browsers or simple "apps", those expensive personal computers will be replaced with cheaper Android tablets and iPads. It'll be those of us who sit behind "the cloud", who develop, who support, who dream, who design, and who want to game who will still be using those big expensive boxes. And chances are, none of us will be interested in ChromeOS.
Don't be surprised in another decade if the "home computer" in its current incarnation is a largely dead concept, replaced instead by "smart" everything and individual user devices (ie phones). Some of us will still be sitting in front of keyboards behind masses of screens.. as we continue to monitor and keep the wheels turning behind the abstract concept that you think of as "the cloud" or "the internet". 

The cloud is, after all, just someone else's computer.