Every once in a while I am reminded how wonderful, and perhaps unlikely, it is that I have a nice, slick, mass market consumer device (apple computer running OS X) that is honest to god UNIX underneath.
And then sad, knowing that this is a brief golden age that will probably pass.
tl;dr a more powerful operating system with worse UI, which is the obvious trade. A less powerful system with a fantastic UI.
Chances are you're going to think that the thing you've been using for years is "good UI", and anything different is "bad UI". I don't doubt that Apple have spent a lot of time designing their UI/UX to work well for most people, but it feels bad to me because it's not what I'm used to.
Either way, you get used to it. I've found programs that make OS X work in a way that's more familar to me, and learned to work with the things I can't change.
As for the cmd button replacing the ctrl button. Never going back to ctrl+<key> for shortcuts. The command button is so much nicer to use, it's your thumb vs you pinky. You don't need to stretch you pinky/move you hand slightly to have access to shortcuts and that's so nice. VERY desturbing at first (I hated it), but never going back. Or maybe I've just gotten used to using my thumb. Ultimately, I think it makes a lot more sense though. No more alt-F4 to close windows, its cmd+Q, no more ctrl+B to bold a text, it's cmd+b, cmd+h minimizes/hides your window etc etc.
Also, what i found infinitely annoying at first and that I really like now on Mac, is the window switching with the keyboard. On Windows, alt+tab switches between all windows (including different instances of same program). on Mac, alt+tab switches between applications and alt+` switches between different instances of a said program. I find that more elegant when you have a lot of windows opened, which i always have.
But all these are tiny little things that add up to something really nice to use, but each individually don't and will never justify OS X being the best OS in the world blah blah. You can probably achieve a similar behavior on windows/Linux/whatever. What's nice about it though is that it's the default setup, no third party app needed/config files. You install it and it makes sense. That's why I'm not switching from OS X anytime soon, it's ergonomic, the UI is sick, power user features everywhere, and a real unix. I see it as a linux that works perfectly out of the box. I remember the first month I spent on OS X (coming from 4 years of linux, 2 yr Ubuntu and 2 yr Arch, using both full desktop env (KDE mostly) and lightweigth window managers), everything works and works well and everything is ergonomic and makes sense. Put simply, a Mac never (or very rarely) gets in the way of your work. And that makes a big difference.
But then, I just like Macs because of how I use my computer. I get it that some might prefer a Linux or Windows, and I respect that. I'm just extremely thankful OS X answers exactly what I expect from a computer.
You mean like with the keyboard? Like SHIFT+TAB? That's on OS X.
Eh. I've used focus-follows-mouse and tend instead to strongly prefer the OS X method of allowing scroll events to percolate to the window under the cursor but retaining keyboard focus on the window it's currently in. This allows me to scroll, say, a web browser on my right-hand monitor, while continuing to type in the left.
Also, don't forget that OS X lets you scroll whatever window your mouse is over, even if it is "blurred" (not in focus). This is the one feature that I miss most whenever I need to use Windows.
The only people here who really seem to like it are the ones using drawing-tablets as their main input.
Never found a good replacement in OSX.
How does the overlappingness make anything faster? If anything it becomes slower, because you need to move windows out of the way, or select them before you can get to the covered part.
> Not everybody wants to use keyboard all the time
In no way necessary. I don't think the WWDC splitscreen demo used keyboard shortcuts at all?
> it is useful to have more than two windows on screen at a time.
Having more than one window on the screen is also not helped by the windows overlapping. I regularly have 2-4 windows side-by-side on the same screen at once without overlapping, and multiple such screens and fullscreen windows on virtual desktops.
It's mainly focused on modifications for the OS X dock which I don't use, but it also does alt drag window movement, and alt + num pad for tiling windows. It costs a a bit of money after a trial period, but it's the most reliable replacement for alt drag I've found so far.
On Unix you have a lot of choice on the (graphical) interface you chose to interact with. So while to one you're currently using might be OK, there's bound to be one that's even better suited to your needs.
https://en.wikipedia.org/wiki/Window_manager
https://en.wikipedia.org/wiki/Desktop_environment
(Also, it's usually trivial to change your keybindings.)
It is like switching from JS to Ruby for me.
We have also seen an incremental approach in the gradual replacement of open APIs/components with closed ones.
> How can operating systems research be relevant when the resulting operating systems are all indistinguishable?
> There was a claim in the late 1970s and early 1980s that Unix had killed operating systems research because no one would try anything else. At the time, I didn’t believe it. Today, I grudgingly accept that the claim may be true (Microsoft notwithstanding).
> A victim of its own success: portability led to ubiquity. That meant architecture didn’t matter, so now there’s only one. Linux is the hot new thing... but it’s just another Unix.
#!/bin/bash
currentProcess=$(xprop -id $(xprop -root | grep -F '_NET_ACTIVE_WINDOW(WINDOW)' | awk '{print $5}') | grep -F '_NET_WM_PID(CARDINAL)' | awk '{print $3}')
state=$(grep 'State:' /proc/$currentProcess/status | awk '{print $2}')
if [[ "$state" == "T" ]]
then
kill -CONT $currentProcess
else
kill -STOP $currentProcess
fiWhere it gets more interesting is using elaborate checkpoint/restore [1] mechanisms for processes to serialize their state (open fds, watches, IPC, etc.) into an image that can then be overlayed, remotely executed, debugged and so forth.
I mean, of course, we have these apps that splay themselves all over global system resources in complex ways, and so we need to track down how they've done that, record it, and then replay it later to restore them. It's necessary.
But imagine a well-behaved classical X11 app, running in something like a VM or Docker container. You execute its launcher, the VM spins up in the background, the app launches in the VM, and the app's X11 client connects to the X11 server you're staring at and displays things.
If we truly want to sandbox apps, wouldn't it make more sense to move more toward a model like that—one where you have a discrete "application backend process" running in a sandbox, and you can just chuck around the sandbox itself willy-nilly (freezing it, thawing it, copying it, etc.) because it has well-defined connection-points to the host?
C/R enables complete persistence by duplicating a process from its image, which can cut initialization and startup times by orders of magnitude, and resume service from an exact position with zero downtime, and not just basic mechanisms like socket superservers.
It lets you back up process state at any arbitrary moment, send it over the network, recalculate and reexec immediately and then thaw again for debugging it from that particular point.
It just helps enable live migrations and zero downtime recoordinations with far less complicated redundancy schemes. This includes containers. On top of it being a universal save/pause button.
Depending on the robustness of the image format, it can be made space efficient through deduplication.
C/R is quite invaluable for any scenario that involves fault tolerance, high response times, tracking down difficult bugs, low startup overhead and so on. I'm surprised why it's still such an obscure technique.
Besides, I don't think the primary use case here is even graphical apps (not that they don't work just fine). What you really want is a local indexed copy of all the execution environment details, like pipes, sockets, timers, ttys, timers, sigmasks, *fds, etc. for infrastructure.
calc
$a = Get-Process calc
$a.ProcessorAffinity = 1
$a.PriorityClass = "Idle"
$b = [Diagnostics.Process]::Start("PowerShell", "-c while (1) {}")
$b.ProcessorAffinity = 1
Ta-da!The bizarre part, when it comes to Flash, is that I've yet to see an actual difference in the application behavior when I throttle down it's CPU usage. I would think that if it normally maxes out 7 of my 8 logical cores, being limited to 40% of one core would make introduce slowness. But I guess that's just an indication of how poorly written the OS X flash client it.
I use cpulimit to force a CPU bound program to slow down. I also use it for programs that always max out the CPU for no good reason, and work fine with less CPU.
Nice won't help with either of those.
It works pretty good, but I have had flash games crash if you limit them too low for a while.
If I'm running a CPU-intensive calculation and don't want to bog down the rest of my system I will use it to give me enough cycles for Sublime Text to work smoothly.
It's easy to call:
ex. 1. http://www.willnolan.com/cputhrottle/cputhrottle.html