Has anyone ever clicked on a random link and been taken to a commercial porn site? To shock sites and stuff hosted in random places, sure, but these are not the ones being threatened with shutdown.

Nonsense. It is true, however, that Windows and Linux use different (overlapping) subsets of the SATA (and SCSI) command sets and, in particular, use very different sequences of commands in common use. If you test heavily with Windows and not with Linux, then you may find that there are code paths in your firmware that Linux uses a lot but which are mostly untested.

The 'tray' that Raymond describes in his second article looks very much like the Shelf from OPENSTEP 4.1, which was released just after Windows 95. I wonder if some of the NeXT people were playing with early betas of Windows 95 and, as their company CEO later quipped, started their photocopiers...

Win32s was released for Windows 3.1, but it just added some win32 APIs, not the UI. The UI was first introduced in the Chicago betas, which were eventually released as Windows 95. NT4 was released shortly afterwards and wasn't a bad OS, but hampered by the lack of plug-and-play support and perpetually having old versions of DirectX.

And, of course, I meant 'to shortcuts' there.

There are a few differences. First, symlinks are a property of the filesystem. This means that the normal filesystem APIs just work with them and you need special APIs for things that care about whether it's a link or not. In contrast, shortcuts are just another kind of file and everything that wants to follow them needs to know what the target is. Second, shortcuts contain a lot more information than just a path: they include the path to the destination file, an icon, the set of command-line arguments to pass, and some other flags. For example, I used to have a load of different shortcuts to the WinQuake (and, later, GLQuake) executable that all had different -game flags, for launching different mods. Many of them also had different icons, if the mod came with its own icon. You can't do that with symlinks.

The closest thing to symlinks on *NIX systems is .desktop files.

Do you live somewhere with an unusually high concentration of stupid people?

I used to walk home through a park. Except on cloudless nights with no moon, you got enough reflected light to be able to see quite clearly across it. Then there some some hysteria about the potential for being attacked (triggered by a flasher, who only exposed himself to people in broad daylight) and they added a row of streetlights along the side of the path. If you stood about 10m from the path, you were completely invisible to someone walking along it, but they were clearly visible to you for their entire trip across the park (as were any potential witnesses on the path). If someone actually wanted to attack people crossing the park, the lights made it a lot easier. It would only take a few seconds to hit someone and drag them out of the visible area.

That's not been true for a very long time. Current MacBook Pro trackpads (i.e. the ones that TFA is about) use one finger tap for left click, two finger tap for right click, two finger drag for scroll, and four-finger drag for some system-wide gestures. The 'Mac mice only have one button' thing is a decade old (and even before that, Macs supported multi-button mice, they just required that UIs be designed to work well with one-button input devices).

Is this really what Windows users consider easy? On a Mac, it depends on the keyboard layout, but for me it's alt-2. A cent symbol is alt-4 (dollar is shift-4). Entering a character with an accent is alt-something for the accent and then the letter that it goes on top of. For example, i-umlaut is option-u then i.

If memorising unicode character numbers is your idea of good HCI, then I really hope I never use a program that you've designed.

Good UIs should be designed to work with a single mouse button, because that means that they'll also work well with a touchscreen. There's nothing wrong with making some things faster with the right mouse button, which is how most Mac applications work. The right mouse button for the context menu was inherited from NeXTSTEP, where the 'normal' menu was a floating version and you could cause a copy of it to pop up wherever the mouse was with the right mouse button (RISC OS took this further and didn't have any kind of menu bar, using the middle mouse button to produce the menu on demand. Using RISC OS with a touchscreen or pen tablet was... interesting, and it only just counted as discoverable because the machines shipped with a mouse with the buttons labelled).

When I'm reading a document, I'll do two-fingered scrolling on the trackpad to navigate. I only use home/end and page up/down when I'm typing, to navigate within the document, and then I already have both of my hands on the keyboard. The function key can be pressed with the knuckle of the little finger of the left hand, so is pretty easy to hit.

Windows-R brings up the run dialog, which will autocomplete program names and used to be the fastest way of launching programs on Windows (it's well over 10 years since I last regularly used Windows, so I don't know if this has changed). Windows-D showed the desktop and Windows-F the system find dialog. All of these were pretty useful, but the key was still quite under-utilised.

Control and alternate already have well-defined meanings. Control is for entering control characters, alternate is for entering alternate characters. OS X uses both. UNIX keyboards used to come with a meta key, but this fell out of use as software was written for PCs without such a key. On OS X, the usage of the command key is inherited from classic MacOS: It's the modifier that you hold for commands. This means that the OS X terminal is the only graphical terminal that I've come across that doesn't suck for copy and paste. On OS X, every single program including the terminal uses command-C for copy and command-V for paste. The terminal is therefore free to use control-C for sending the character that they terminal recognises for SIGINT. Windows overloaded the alternate key for opening menus, which meant that it is no longer a convenient key if you need to enter non-ASCII characters (for example, a Euro symbol or a letter with an accent, which are both easy to enter on a Mac). Most desktop environments for Linux inherited a load of bad UI design from Windows before adding their own mistakes.

The difference between persistent and temporary storage is important. Being able to have 128GB of RAM in a laptop that consumes no power when not being read or written would be a huge win (one of the reasons phones have limited RAM is that DRAM draws power all the time) would be very nice.