The Linux Experiment

[Page 3] The Results

The Results

It's true to say that I only used Mint for a relatively short amount of time, so in no way did I, or could I, experience all of the features and issues of Linux. Keep in mind that the previous page was a condensed summary of some highlights and lowlights from my Linux experience. It could be argued that I made avoidable mistakes, or didn't try hard enough, or quite simply that I'm stupid. But overall, I believe I examined the good, and the bad, of Linux, from the point of view of a long-time Windows user and Linux noob, with a reasonable amount of effort - as was the stated goal of the Linux Experiment.

First Impressions

User Interface: Linux Mint Cinnamon seems to be an excellent choice for any Windows user wanting to attempt to transition to Linux. The developers have invested a great deal of effort into making the interface and workflow as 'Windows-like' as possible, so that a life-long Windows user can confidently begin navigating around the OS from the minute it boots up.

Performance: Linux Mint Cinnamon was very responsive on my system, although I am running higher-end hardware on a fairly new machine, so that's not really surprising. Still, I was very impressed with the way that Linux, out of the box and with no tweaking, booted up at least as fast as a tweaked Windows 10 install with Fast Startup enabled. Same deal with general desktop usage: everything was snappy in Linux Mint, and aside from the mouse scrollwheel sensitivity being a bit too high, the default configuration is at least as responsive and easy to use as my Windows 10 desktop. This is even more impressive given I didn't have to manually install drivers for all of my key components to get it this way, Linux took care of all of that, bar the graphics driver.

Stability & Bugs: Linux Mint Cinnamon feels quite refined in terms of system stability and predictability, and I didn't experience any serious bugs or unexpected behaviour. I even did some impromptu partition management on the Linux system partition, and despite all warnings that this is very risky in a dual boot environment, Linux and GParted handled it superbly, fixing what looked like some resulting problems without any user input. The most noticeable bug I encountered is visible on these pages: some screenshots taken using the PrtScn method in Linux result in minor glitches being shown. These glitches did not appear on-screen while taking the shot, and it's likely they are due to the Nvidia driver. Speaking of which, one of the biggest usability issues - the secure boot-related Nvidia driver problems - wasn't a bug, but it does represent what could be a deal-breaking hurdle for some users, especially since even when I spent some time researching it on Google, I didn't find the right answer; only a random user comment on the download page for the Nvidia Settings utility, which luckily caught my attention, gave me the hunch to check my UEFI BIOS settings.

First impressions are very important, and Linux Mint Cinnamon passed this test fairly well. The Nvidia Driver stumbling block, although serious, can be mitigated simply by having Mint provide more detailed instructions to the user regarding Secure Boot, e.g., a prompt urging the user to check their UEFI BIOS options to ensure all boot-related settings hand off to Linux correctly. At the moment, it gives a misleading impression that it can handle it all without any need for the user to act.

All of this was just scratching the surface. The real issues became apparent as I used the OS in earnest. I'll cut through the petty annoyances, and discuss the key areas of concern.

Linux's Problem is Terminal

A modern Graphical User Interface (GUI)-based operating system like Windows allows the average user to do virtually anything, even relatively complex system administration tasks, such as Registry editing, installing/uninstalling drivers, setting system policies, viewing error logs, etc. via a graphical point-and-click (or touch) interface. Knowledge of MS-DOS or PowerShell commands is handy, but not necessary in the slightest. Furthermore, the Registry, DOS, or Powershell commands are fairly consistent across all editions of the same version of Windows (i.e., Home, Professional and Server editions of Windows 10), and even across newer or older versions (i.e., across Windows 10, 8.1/8 and 7). Thus users needn't worry about having to modify these commands for their needs, a simple cut and paste of the relevant commands from a guide for example usually does the trick.

Linux doesn't quite work this way. My brief experience made me realise that even moderately simple tasks, such as customising the interface with a theme, can see the user suddenly immersed in a Command Line Interface (CLI), via the Linux Terminal known as Bash. As a form of direct access to the Linux kernel, the terminal is an incredibly powerful tool in the hands of an experienced and knowledgeable user. For a long-time Windows user and Linux noob like me however, using the terminal feels like I've regressed all the way back to the 1980s. In those early days, I used AmigaDOS, then MS-DOS, then learned the fundamentals of programming with simple languages like BASIC and TurboPASCAL. Command line interfaces were all I had to interact with the operating system in any detail, so I learned them, and I used them. But I never liked them. As much as it gave me a cool movie hacker vibe to type out a bunch of commands just to move a single file from one directory to another in MS-DOS for example, I was so thankful when we eventually got a simple point-and-click Windows Explorer UI in the 1990s.

Linux Mint sends a mixed message to users. There is a very slick GUI at first glance, and very obvious attempts to align the desktop environment with common Windows usage conventions, and to even provide key Linux functionality in the form of GUI-driven utilities, such as the Update Manager; a tacit admission of the tediousness of using Bash commands to do these common tasks. But for all of their efforts, the Mint developers can't alter the fact that Linux is fundamentally a Terminal-orientated OS. This shouldn't be a surprise given it was originally derived from UNIX, which is a purely CLI-based operating system. Linux's primary user base is heavily skewed towards people who are experienced tech enthusiasts and/or IT professionals. These are the sorts of people who are comfortable with learning to work in a CLI environment, and some would be using a CLI on a daily basis as part of their job anyway. Thus current Linux users and developers wouldn't see a reliance on terminal commands as being any sort of hindrance; quite the opposite in fact.

Logically however, I stand firm on this point: a home PC user should not need to use terminal commands to do any standard task in a desktop OS. There's a reason why the GUI was invented, and it's because it is a user friendly, and often far more efficient, way to manage a range of tasks. Unfortunately, the standard response to this observation from Linux enthusiasts is some variation of "Linux isn't Windows", "Learn Bash or GTFO", or some other snarky comment. This doesn't change the fact that the average computer user is not a tech enthusiast, and so doesn't want to learn a bunch of Bash commands, especially when there are other options. Consumers are after products that are less demanding in terms of physical and mental effort, not more. I was hoping that Linux Mint Cinnamon, by virtue of its widespread reputation as the most user friendly of Linux distros, would accommodate this group. It tries, but in my opinion, fails, because of the poison pill at the heart of Linux: the appropriately-named terminal.

To be fair, I totally understand, and even support to some extent, Linux users who say that their OS shouldn't be dumbed down just to suit the average Windows user. There is truth to the oft-used retort that Linux isn't Windows, and it shouldn't have to be. Linux is an alternative to Windows, not a clone of it. But that basically locks Linux into its single digit share of desktop users. Market share is a question of popularity, of pandering to the masses who may be less intellectually capable of, or simply just not interested in, understanding complex technology. These sorts of people do not want to put effort into learning a long list of terminal commands, the correct syntax for their usage, and having to research methods to install each new game or program. Speaking from experience, I know that a lot of people don't even want to delve into using a simple GUI-based OS like Windows to their advantage, so Linux has a snowball's chance in hell. As more and more people turn to using gaming consoles, mobile devices, and the built-in operating systems in consumer electronics - which are all deliberately designed to minimize user thought and effort even more - they'll garner even greater market share. Software developers are attracted to the platforms with the greatest market share. The lower the market share, the less a platform receives developer attention. Which in turn makes it even more of a niche platform, because of the time and effort required to implement workarounds and hacks - via a CLI - to make things run on it.

There is one Linux distribution that attempts to be an actual Windows clone, called ReactOS. It doesn't just try to look like Windows, it actually allows the use of Windows apps and drivers, which is particularly important for gamers. But a quick look at its most recent major release, ReactOS 4.11, shows that it's got quite a way to go.

A Question of Motivation

Each time I ran into an issue with Linux, I found myself asking: "Why am I doing this?" Why was I bothering to go through all of the hassle, when I already had a perfectly functional OS that I had a high level of proficiency with, could do everything I wanted with minimal effort, was already configured to my needs, and which I'd even managed to legally obtain for free? I suspect many a Windows user who tries to transition to Linux will ask themselves the same question over and over again. You will have to answer this question for yourself, as everyone has a different answer. But I'll attempt to cover the two most common reasons for switching to Linux below.

Distrust of Corporations

Some people seem to distrust, even outright hate, large corporations, so a major reason people give for why they want to switch to Linux is because it isn't owned by a large company like Microsoft, or Apple, or Google. Unfortunately, we're far too short on space in this article for me to examine the validity of any criticisms of capitalism and profit-seeking behaviour. The best and briefest argument I can make as to why I don't hold this point of view is that I too don't trust corporations, but I trust corporations to do what they do best, which is to pander to consumers in order to maximize profits. That means they're just as afraid of us as we are of them. So when people panic about, say, losing the digital goods they've purchased, such as games, music and movies, from platforms like Steam, iTunes and the Google Play store, I like to remind them that any company that suddenly removes any significant portion of a customer's digital library will essentially be cutting their own throat; massive negative publicity will ensure that the majority of consumers will never trust that platform again.

This issue is further complicated by allegations of spying against Microsoft, with its encrypted diagnostic data flows within Windows 10, as briefly mentioned on the first page. So a specific reason why many Windows 7 users may wish to switch to Linux is to steer clear of this spying. Once again, this is not a point of view to which I wholeheartedly subscribe. I like my privacy, but I know that there are many layers to privacy. Even if you can fully secure your OS, you may still have a huge privacy backdoor open thanks to Intel's Management Engine or AMD's TrustZone opaque 'security co-processors' quietly churning along in the background on your PC 24/7, even when the power is off. See this page for more discussion of the Intel Management Engine for example. For me, the clear and present danger from continuing to use Windows comes from witnessing the ongoing incompetence demonstrated by Microsoft, first with key decisions relating to the evolution of Windows, such as the Windows 8 debacle, and now with Windows 10. Illogical changes, constant backflips, dubious statements, quality control issues. These are all indicative of a company without a clear vision, one which hasn't worked out a viable monetization strategy for a desktop PC OS in the mobile era, and so is constantly reconfiguring the OS to meet their needs, and cutting corners to reduce costs - like crowdsourcing the testing of critical OS updates. So it's not so much that I fear Microsoft's intentions, it's that their incompetence puts my system, and my data, at risk.

So it boils down to exactly how much you fear, hate, or mistrust Microsoft, or any other company, as to how far you go in putting up with Linux's irritations and its learning curve.

Technical Merits

Another common argument put forward in favour of Linux is that it is better purely based on its technical merits.

Security: Linux is often presented as being more secure than Windows, but this is debatable. It seems intuitively plausible that Linux, the OS of choice for many servers which are targets to all sorts of attacks, would be the most secure. But this article for example presents a comparison of Linux vs. Windows security out of the box, and there is no clear-cut conclusion. This is because security is a function of many, many variables. Let's take one specific example that frustrated me no end in Mint: the more onerous privilege escalation requirements for Linux user accounts, namely having to type in the (long and complex) Superuser password each and every time you need to make a system-intrusive change. This is presented as being a better security feature than Windows UAC's default behaviour of logging in as the Administrator once, and then clicking OK to UAC prompts thereafter to escalate privileges. Some Linux supporters claim that the former is more secure than the latter, citing scenarios such as a colleague accessing your PC to install harmful software while you step away to get a coffee. A counter-argument is that firstly, the UAC in Windows can be set to a higher level by default to achieve the same result as Linux, and secondly, a more efficient solution is to get users into the habit of pressing WINDOWS+L to lock their workstation each time they're away from it for more than a few seconds. Faced with having to type in long complex passwords many times per session, or pressing WINDOWS+L a few times a day, most users quickly change their habits.

I don't pretend to know conclusively whether Windows is more secure than Linux, or vice versa, but some research quickly reveals that there is no conclusive answer, even among the experts. It all depends on how secure you need to be, and the extent to which you've lock down either OS, sacrificing convenience in return for greater security.

Performance: Objective tests of major Linux distros, such as those provided in this article, tend to favour Linux over Windows 10 in terms of raw performance. This, along with the typically zero cost of licensing, is why Linux is the OS of choice for manufacturers who make low-powered and/or energy-conscious devices. However, these tests of Linux's raw power are typically moot for desktop users, because the most resource-intensive applications the average desktop PC user employs are games, and there are many reasons why Linux cannot be as fast as Windows for most games. The underlying reason is that developers usually make PC games with Windows in mind, which is perfectly logical given 90%+ of their PC customers will likely be using Windows. This means that any other OS will have to translate any custom calls into their own native format, which inevitably adds some level of processing overhead. Older games may still perform quite well, but for gamers looking to eke out as many FPS as they possibly can in the latest strenuous games, not to mention for hassle-free compatiblity with any game, new or old, Windows 10 is preferred.

Customization: By virtue of its open source nature, Linux encourages choice far more than Windows. The presence of over 200 versions of Linux vs. a few Windows 10 variations, in and of itself, proves this point. But equally, Windows, with its proprietary nature and far larger user base, is more enticing to software developers who can be sure of greater monetary returns if they create a customization utility for Windows. It's difficult to come to any clear conclusion once again, because it all depends on your needs and your capabilities. Anyone can quite easily customize Windows 10 using a tool like WindowBlinds, whereas I had difficulty even getting a common theme to function correctly in Linux Mint Cinnamon.

There is no outright winner here. Switching to Linux may stop Microsoft, Google, Apple, or the government, from collecting any personal data from your PC at the OS level. But there's still the very real possibility that your browser, your ISP, your cloud storage provider, and/or your CPU, are doing that anyway, even under Linux. In terms of general security, it seems unclear. Out of the box Linux has a more secure Superuser feature for example, while Windows comes with a highly-rated antivirus built-in and running in the background by default. If you've got a less powerful system and need a performance boost, Linux will restore some snap, crackle and pop to it. But the flip side of that coin is that you will undoubtedly have compatibility and/or performance issues with some software, at some point, because "Linux isn't Windows", remember? It's not as simple a decision as fans of either OS want to make out.

Free & Open Source: The Best Reason to Switch to Linux?

As far as reasons go for switching to Linux, this is the mother of them all: Linux is based on the Free and Open Source Software (FOSS) model. Linux users will cite FOSS as being superior to proprietary (closed) models, and the best reason to switch to Linux, as though it's a self-evident truth. But is FOSS really the best approach? There's certainly much to like about the FOSS ideal: talented software developers generously providing their time to work co-operatively in order to deliver innovative software, written in non-proprietary code, which can be openly distributed to everyone for free, whether to enjoy or modify, in line with the permissive GNU General Public License. It's the path towards a truly egalitarian society. However, the closer you look at the FOSS model, the less sensible and sustainable it appears to be.

Excuse my cynicism, but I strongly suspect that the most attractive part of FOSS for most end users is the 'Free' part. If we retained every other aspect of the model, and changed the Free part to Expensive, support for this new "EOSS" model would rapidly evaporate. It's no secret that most people don't like paying for something if they don't have to. Therefore, ideologies that give people what they want for free are unquestioningly and zealously championed. Piracy is one example; many years ago I did a major study of PC Game Piracy, and the unsurprising conclusion, based on a large amount of data, was that people support piracy because it's easy to do and gives them what they want for free. All sorts of detrimental outcomes from piracy, such as the mass-piracy of a DRM-Free charity game, were justified or brushed aside, in the name of this conveniently self-serving ideology.

Logically, there is no such thing as free software. Everything has a cost, and someone must cover that cost. Software developers can't pay their food, clothing, energy and accommodation bills with feelings of self-satisfaction or community gratitude. So where does the money come from? For many FOSS developers, the income comes from their day job, working for regular companies, which in effect subsidises their open source work.

One of the best-known FOSS projects is the Firefox browser by Mozilla. What many people don't know is that Firefox has been almost entirely dependent on Google for the bulk of its funding from 2004 to 2014, averaging around 85-95% of its total revenue. Mozilla's publicly available revenue figures from 2013 for example showed that 97% of Firefox's $314m in revenue that year was from royalties, with over a quarter of a billion dollars from Google alone. Mozilla briefly switched to Yahoo as the major sponsor in 2014, but returned to Google in late 2017. While the recent proportions of Google funding aren't provided, Mozilla made a total of $562m in revenue in 2017, and it seems quite likely that, just as before, the vast majority was from corporate sources.

Let's look at Linux Mint, which at one point reached the number one position in the Linux Distro Rankings, and is currently in third place. It receives its revenue from an unspecified mix of advertising, sponsorship, and yes, donations, data on which is publicly released: in 2018, these totalled $146,000. That's it. One of the top Linux distros receives total donations from around the world which are just enough to cover the salary of a single software developer or IT manager. This leads us to assume that the bulk of funds come from corporations in the form of tech support services or direct corporate sponsorship.

Ubuntu, the main distro upon which the Mint desktop environment was built, supplements direct corporate funding and payment for support services by incorporating ads, Amazon affiliate links, and collecting personal information from users in their version of Linux.

This story tells a tale common to FOSS developers - the realisation that while consumers and corporations love FOSS because it's free, it doesn't lead anywhere for you as a developer if you try to make it pay. Of the many people who loved and used his free tool, none supported the paid 'Pro' version of it:

Zero people have purchased Toodles Pro to date... I just wanted to capture some of the value that my work was bringing to commercial entities... I gave up.

There are initiatives like Tidelift, which is trying to develop a working business model that actually helps FOSS developers get paid for their work, but it's not easy. In many ways, FOSS can totally devalue the work invested in software development, flogging it off cheaply to the only people willing to pay for it: the corporations.

Forget about the money for a moment. What about the freedoms that FOSS provides? This too can have dubious outcomes, and Linux provides the best example. The main body that provides guidance and support to Linux projects and the open source community is The Linux Foundation, but it has no power to enforce any overarching Linux strategy. So while the freedoms of Linux may attract developers, the end result is substantial fragmentation, and potentially highly inefficient use of resources, as efforts are split between several competing pieces of software trying to do fundamentally the same job in dozens of sometimes only slightly different ways. It seems that any time Linux developers disagree on a certain aspect, they split off and create a new distro. That's how Mint was created: developers who didn't want to continue working on the GNOME desktop environment split off and made Mint, the Cinnamon version of which I've been using. As a consequence, there are over 200 distributions of Linux at the moment. While ardent Linux supporters will tell you that this wealth of choice is actually a huge plus for Linux, some, like Linus Torvalds, disagree:

I still wish we were better at having a standardized desktop that goes across all the distributions. It's not a kernel issue. It's more of a personal annoyance how the fragmentation of the different vendors have, I think, held the desktop back a bit.

Please don't get me wrong. I genuinely admire the noble ideals behind FOSS, so much so that I changed my site to align more closely with the free open source approach. However, the reality is that the FOSS model has admirable, but unrealistic, aspirations, and users rally behind it mainly because it gives them free software and the illusion of anti-corporatism, while corporations gladly use it to reduce costs, increase profits, and indirectly assert their influence.

The truth is that most of the freedoms enshrined in FOSS are not enforceable in any way, they're more like wishes, and they don't necessarily result in efficient or equitable outcomes anyway. Those that are enforceable, such as the GNU GPL, are entirely reliant on the traditional copyright system, and are not a replacement for it. The end result of the FOSS ideology is that developers release their software and code for free, but then have to go, cap in hand, to corporations to receive the vast bulk of their funding. This merely masks the true source of the revenue on which the FOSS model operates: traditional companies, which earn their money the traditional way. After a fair bit of research on this, I'm left with the inescapable conclusion that FOSS is a wishful ideology completely dependent on the continued patronage of profit-maximising corporations, supported by self-interested consumers who enjoy free software, carried out by well-intentioned software developers who basically have only their faith that it may lead to a reasonable income and beneficial outcomes.

In short, FOSS has more in common with a religion than a business model.

Should You Switch to Linux?

The obvious question is whether it's worth your time to switch to Linux. I don't have a one-size-fits-all answer. What I've tried to do in this article is give you food for thought, so that you can hopefully make an informed decision from my experiences and musings. I can say with confidence that it's at least worth your while to try Linux for yourself using a LiveCD. Quick, easy and harmless.

As for my decision, I won't keep you in suspense: I decided to remain with Windows 10 for now. I'm grateful that Linux exists as a competitor to Windows, and having learned the basics of Linux Mint, I'm now much better prepared with a viable alternative should Windows 10 burst into flames. The main reasons I had considered switching to Linux in the first place are that I'm uncomfortable with Microsoft storing any of my personal data, harvested as part of the "telemetry" in Windows 10, on their servers, as it exposes it to potentially being hacked; and I have concerns about the impact of Microsoft's continued incompetence with patches and major updates for Windows 10 on my system.

Having used Linux Mint Cinnamon, I can honestly say that it's much better than I thought, and dead simple for any Windows user to pick up and start using without any real training or research. But that isn't enough to get me to move over to Linux. The more I used Linux, the more I realised that while basic Linux usage is easy; using Linux in an advanced manner requires a substantial investment in time and effort to learn its intricacies. It's taken me at least 20 years to get to a level of intuitive expertise with Windows that can see me troubleshoot any issue relatively quickly, do anything I want on my system, and have it working with perfect stability, day in, day out. Retraining myself in Linux to achieve the same thing seems like a silly waste of time just to avoid being at the mercy of Microsoft's incompetence. The Linux terminal and package dependency issues clinched the deal.

My opinion is that Linux will gain little, if any, market share on the desktop; there will never be a Year of Linux. In the long run, I see entirely GUI-based, touch-centric OSes like iOS and Android being dominant anyway, so spending time mastering Linux isn't going to benefit most people in the future. Anyone who thinks that the average user will be typing out commands in a Linux terminal to get things done 5 or 10 years from now is, I would suggest, delusional. The Linux kernel will always be there, in various devices and powering servers. But as a desktop OS, it doesn't beat Windows 10 sufficiently on technical merits alone, and its development is fundamentally flawed through inefficient fragmentation, a reliance on terminal, and ultimately, the well-intentioned but unsustainable FOSS model.

That brings the Linux Experiment to a close, and I'm sure it'll be as warmly received by Linux fans as a BLT at a vegan picnic. The results are largely subjective, but I've made every effort to provide clear rationale and supporting evidence for my conclusions. If you find any factual inaccuracies, or just want to give me constructive feedback, email me. Keep in mind that I can't provide any technical support or personalised advice on whether you should switch to Linux. Otherwise, I hope this article helps spur some genuine discussion on this topic, and please keep it civil.

Until next time, take care!

Go to Page... Page 1: Introduction Page 2: The Experiment Page 3: The Results