I never said, nor even suggested that you were ignorant. Your opinion just seemed to be based on outdated experiences. So some feedback, for what it's worth:

1) I have Kubuntu running reliably on 3 very different setups. If win8 is working well for you on your setup and Kubuntu isn't, then it's time to start blaming the hardware manufacturer for not giving a flying fsck about the users and only putting effort into windows drivers. You do remember, after all, that linux drivers are (mostly) done by unpaid people who do so just because they can and care to? Not to say that I don't care about your plight -- PLEASE do email your hardware manufacturers and implore them to solve the situation. It's them that are depriving you of the freedom to use your device as you see fit. In that vein though, I've learned (the hard way) not to buy hardware which isn't mainstream and already well-supported by the FOSS community simply because I don't want to be in that situation. Not to say that I get it right all the time :/
2) I've installed Kubuntu (and other debian derivatives) with zero issue under VirtualBox. Multiple times. Again, if your virtual machine provider doesn't provide for an environment as supported by the FOSS community, please DO complain loudly to them. You'll be doing everyone a favour. Still, you need to know that VM+*buntu == win, so if that's not your experience, swap out your VM.
3) Every experience I've had under OSX has been one of frustration. To the point that, in addition to my existing base rage against the cost of crapple products for what you get, hardware-wise, I'll go so far as to do whatever I can to discourage someone from getting a Mac simply because the user experience that I've had has been one of terrible frustration. "Close" window titlebar buttons which don't actually close the application. "Maximise" buttons which make the window arbitrarily bigger, but not to the size of the screen. A "POSIX" system which has the terminal buried under layers of tricky-to-traverse menus (layered menus are one of the greatest UX failures ever, imo), and a system which has the intelligence to use the 'file' function to figure out that a .xlsx file is a zipped collection of XML and dump that out to the desktop when a user without MSO double-clicks it, but which provides no feedback about that eventuality. This is UI failure. When the ideal target audience for a Mac (inexperienced user, given laptop for birthday, not a power-user on any platform whatsoever) can't figure out why her mail attachment "won't open" (when it's silenty unzipping to the desktop with every double-click), it's time to give the whole deal another thought. Windows could have done better. A decent Linux distro could have done better, though I'm sure a shitty one wouldn't have.

I really try to be objective and I'm a big proponent of "a tool for a task". But, at no point in the time that I've used OSX, have I ever thought "I could carry on using this", or "Other vendors could learn from this". Multiple times, the phrase "HULK SMASH" has been predominant. YMMV, of course, but the Linux desktop is way less frustrating and way more stable and responsive, as my pre-schooler would attest to -- if he even cared. And he doesn't. And that's the point. It gets out of the way so he can do the stuff he wants to.

The problem was never that Ubuntu became popular. Geeks everywhere rejoiced!

The problem was simply the Gnome3/Unity fallout which left a lot of users with no easily accessible default desktop. Each has interesting ideas and strengths but neither are the stalwart that Gnome2 was. Anyone who was already on the KDE track with Kubuntu wasn't bothered though... until Canonical dropped official support for Kubuntu.

Personally, I still run Kubuntu. KDE plasma has evolved from a bloated pig into something pretty and acceptably fast. There are still quirks, but they're less than what I have to deal with on any other platform. Kubuntu is my "daily driver", as it were, with a dual-boot to win8 for games (those not found under Linux; a number which is diminishing) and when I really want to use Visual Studio (usually work reasons). I used Mint for a while, but learned the hard way that a distro which recommends a re-installation over an upgrade is a bad choice for a desktop which is expected to be alive no-matter-what.

You obviously haven't visited Penguin land in some time. I suggest you go find a friendly KDE-based distro (like Kubuntu, but there are others) and test your hypothesis again. You may find that you get a better experience at zero cost.

Funny you should put it that way.

I just installed a laptop for a 3-year-old who likes to play games out of my Steam library. It's running Kubuntu and he has no problems with it. It boots fast, plays his games well and generally gets out of the way whilst he enjoys himself. I'd hazard a guess that he's below the average user, he's running it as a desktop and he's made ZERO tradeoffs. And neither did I.

Sure thing; just balance that with the obviousness that disgruntled users == loss of revenue, either directly or through migration away from your products. There are quite a few Android spinoffs without the googleness in them, so the googleness is quite well baked out of the OS. Not saying it's perfect, but the sheer structure of the community and the opensource nature of the code makes it a more difficult target for the spybies. Then again, nothing is unhackable, nothing completely safe. So I guess I go back to the mantra "pick the battles worth fighting and live your life there rest of the time" (:

I dunno, perhaps I'm too much of an optimist, but:

1) I think google only wants to break privacy if it increases ad revenue; I think they'd rather "stick it to the man"
2) the latest Apple snafu with fapgate is really just dumbness, not nefariousness (if that's even a word). I mean, no lockout on incorrect password == brute force win, duh.

Perhaps they are t3h 3v0lz. Perhaps not. I prefer to "occam's razor" shit and take the simpler solution, albeit that I recognise that I may not be right; I just don't want to live constantly looking over my shoulder and I have to admit that a resounding metric buttload of conspiracy shit is exactly that -- shit. Some may be truth, but it deserves the same scrutiny we place on the "truths" we're told every day if we want to stay sane. Just my /2c.

Like I say, I'm not disputing that there are a bunch of nefarious fucks trying to run the world; just that, if iOS and Android encryption are bunk, they either went to a great deal of effort to make them resource-intensive or they just plain hired a bunch of PHB's, pointed them at Scratch and said "go code encryption, 'cos you can!". Because seriously, Apple's planned obsolesence is working and I had to turn off encryption on my s3 after I started getting the urge to smash it.

Cool, I don't know about iOS for sure, except that iOS 8 fucking lags on an iphone 4s, so I have to assume it's doing *something*.

In addition, enabling encryption on my i9300 (android, of course), led to tangible lag in device usage. If you're going to make "tissue paper" encryption, you'd at least omit the lag, surely? Not that I'm fully refuting your claim -- just saying that if it's true, someone went to a lot of effort to waste cpu cycles so it seems as if there's something happening.

the other day. Here's what happened:
1) Performance sucked ass, despite reports to the contrary (i9300 -- I know it's no G3, but hey, it should damn well be enough, at quad-1.2 gHz with a gig of RAM)
2) My TWRP restore didn't include my home partition so I lost all data on there. Sucks to be me.

I'd welcome this if it didn't come at the massive lag that I experienced on a device which is normally quite spritely. I get that encryption doesn't come for free, but adding 1-3 seconds lag to every tap is not, in my book, worth it. I'd rather just use the android device manager to remote wipe if I lose my phone.

Because software is never finished. There's always an optimisation or a bug to fix or a feature which falls in line with the ethos of the project.

No updates == no interest from the developers and that's a bad place to be in. I'd rather use software which is actively maintained (and has some issues) that something which is robust but stagnant.

Change is the only constant.

You seem to think that a microSD slot is an unambiguously bad feature. You seem to have a list of downsides which are quite flimsy:

1) I replaced the sim slot card on my old S1 the other day (by myself, because it's actually not that difficult to do and cheap to do as well) because it had been abused and broken -- no longer reading a sim. The SAME BOARD hosted the microSD slot. It's not really a bulk issue then, is it? It's a tiny 0.5x(square dimensions of sd card) mm cubed "bulk".
2) It was on the same board. I bet it must have added around 0.5c cost.
3) The sd card slot didn't die -- the sim card slot did. Reliability of the sim slot seems to be more pertinent.
4) No-one is going to try to wedge battery in there. Not at the space it takes. Not in the place it's at. A good place to put more battery would be where that stupid fingerprint scanner is or where the NFC chip is since contactless payment is a giant fail and I've yet to see a truly innovative use for NFC. I certainly don't miss it when I swap out my S3 battery for a higher-capacity one which has dropped the NFC chip. But that is my preference, of course.
5) Depends on the model -- all of the Samsung phones I've seen (s1,s2,s3,s4, note2) have the sd card slot behind the back cover, Pretty sure the HTCs I've had contact with were the same, but I could be wrong as I didn't deal with them that long. Indeed, I don't know offhand of any modern phones which have the slot exposed like they used to, though I'm sure some exist. The point is, they don't have to.
6) The majority of Apple's devices are "third party hardware". Again, issues with sim card slot, not sd card slot. That same phone had an issue with the main board earlier on (replaced under warrantee) -- but never with the sd card slot. Odd hey?
7) True, consumers tend to choose the cheapest. Though most who are savvy enough to buy a card, shop around for one which says "fast". Or they ask a friend, like they did when they were buying other technical stuff.
8) How so? If you have a *nix-style filesystem (as do Androids and I'm pretty sure so do iPhones), then the card is mounted in somewhere into another folder. Meaning that if nothing is mounted, the folder is just a folder on the original filesystem. How is that any more complicated than any other folder on the filesystem? And if it's mounted, the fact that it is is transparent to anything casually observing the fs.
9) Preference is a fair point, but not a downside. If you prefer smaller storage and wireless sync, go ahead. Unless you're in a really first-world area though, storage of larger files becomes an issue you have to face as you can't just stream everything all the time. So it's valid to have a preference for more storage and, beyond that, the ability to expand that storage as requirements are raised, in accordance with available finances.
10) Only in the same way that USB drives, SSDs, DVDs... indeed *all* storage media are "modern day floppy disks". In the way that they are all means for storing data with variable amounts of portability. I think we do need storage media for our data, but I could be way out on a limb here. Behind the cloud is some kind of storage. Also it's quite rare to find people popping their sd cards in and out of their devices on a regular basis, so the "floppy" analogy is a bit flawed on that point and immediate access to the card is also not a priority.

I'm particularly amused that you instruct the prior poster to "stop trying to tell everyone that they are stupid because they don't care to do things just like you would" after having basically done just that. Some people want storage. Some people prefer wireless sync. It's all about preference, but you've simultaneously told another poster that their preference is stupid and that they shouldn't tell anyone else that same message.

On one point we can agree though: that poster should probably get out of the Apple camp and find a device where the makers are more interested in the needs of the user than in serving their own agendas and trying to crowd the world's landfills with discarded, fully-functioning electronic waste after collaring their obedient sheep into buying the next iteration of the same thing they did last year, only this time with most of the features they shot down in competitor products two years before after realising that their blind followers were starting to wake up and smell the coffee.

And I'm not trying to spew vitriol about Windows or services in that platform, though I'm sure you might expect that here and I might even have appeared to be doing just that (: I was just answering your question (:
I've come to understand that all tools have their rough edges. I make my living from writing software for the Windows stack and most likely will continue to do so for many years to come.

The biggest issue I have with the Windows service model is that certain services are run in the same shared process, which means that if you want to take one down, they all come tumbling down.
Secondary to that is the flakiness of the install/uninstall process -- I've found it far too easy to get the services subsystem into a dodgy state where a service is scheduled for deletion, but not actually deleted until you restart, even though it's no longer running or in use. Which means that you can't re-install a newer version of the service until you reboot because your new version shares the same name.
Tertiary to that is the fun and games that you have to endure to be able to run an application as a service but also run it once-off for debugging / development. Hence the birth of projects like this: http://www.nuget.org/packages/... (full disclosure: I wrote it and (a) I'm sure it's not a unique solution and (b) there are probably better ways of doing it but (c) I found I needed to do this far too often). Much of this hinges around how messages (start/stop/pause/resume etc) are sent to the service -- and these are also (IMO) far less elegant than *nix signals. It's also this architecture which makes a flaky service gain the ability to cause havoc with your system by not responding to requests properly, etc.
In addition, programming a service for install / uninstall is a mission in and of itself -- even the "standardised" .net way is well unintuitive. Again, this is solved in my shell, using http://www.nuget.org/packages/... (again, full disclosure, written by me), but basically involves telling the system where to find entry points in your service code; in other words, forget the standard main(), there are other entry points that the Windows service system uses. Coming from another platform, you'd expect services to be easier to write -- they're just regular, daemonized processes launched from an init script (which you can find a template for easily enough).

All of that doesn't mean that I could have done better or that it's absolute shit. It's just that it takes time to figure this stuff out -- many burned fingers before you start getting it right. And I still get the feeling (oodles of services later) that I'm missing something. Because I probably am.

I sincerely hope that systemd doesn't foist this same mess on us. I haven't investigated it enough to know much about it, but the facts that I have read (primarily who the dev team are and my experience with their prior fuckups and most especially how they like to shift blame after royally fisting someone else but also how they don't give two shits about user problems or reinventing stuff that didn't need reinventing) don't give me that warm fuzzy feeling.
PulseAudio is a great example: pretends (poorly) to be ALSA and then the devs blame userland software when audio gets choppy (even when audio wasn't choppy on ALSA). It also poorly implements features like simultaneous output to multiple physical devices and devs don't take bug reports on that shit -- I tried and was shut down for using a feature I apparently wasn't supposed to be using (ie, simultaneous output). They'd rather leave broken shit in place than fix or remove it.

I'll be honest that I don't have a broad range of experience with libraries. I've used a couple of html-to-pdf implementations and PDFSharp. The licensing for PDFSharp is very permissive, support can be paid for if required and the library is quite fast. As an aside, it has a cousin, MigraDoc, which produces abstract documents which you can finalise to Office formats, if you need that too.

IMO, there is no perfect tool, but PDFSharp has served me well.

Has this got to do with the article? Or the OP?

Indeed, what the actual fuck are you smoking? And where can I get it?