Turned Off¶
1058 words on Mac OS X
Joel Spolsky wrote on Windows’ shutdown options (with a classy screenshot exhibiting a Windows look that’s probably called ‘Zune brown’) and the processes in Microsoft turning such a simple feature into a confusing nightmare. He also linked to a more detailed description of how such quality features are made.
To see the other side of the software world, there’s also a text on how Apple’s shutdown options came to be. On the Mac you only get the options to shut down, restart or sleep in a fairly straightforward way.
And the author argues that having even less options would be preferable. In theory that sounds like a good idea. But in practice, I am pretty sure that the current setup is a good compromise. Let’s just look at the three options:
- Sleep
- In principle the computer goes to sleep automatically when it is idle, so there shouldn’t be a need for the ‘Sleep’ feature. In reality computers sometimes don’t go to sleep when idle, because some process keeps accessing the disk or so. And from power usage considerations to the fact that you can set up the computer to lock the screen when waking from sleep, there may be many reasons to want the computer to sleep now.
- Restart
- In principle there is no need to restart the computer. Because it’s stable and it ‘just works’. And if the computer’s system software is updated in a way that requires a restart, the updater will provide a handy restart button. In reality however, things can and do go wrong from time to time and you just need to restart the machine anyway. Be it because your font caches needed cleaning or some comparable technical nonsense. Or because OS X has this habit of littering your hard drive with swapfiles that you may want to clean out when your hard drive is well filled. And even if space is not the problem, I keep having the impression that once the computer’s memory has been filled and you bring the virtual memory system into serious unbalance by launching some memory hog application (fun way to try this would be to open a large image in GraphicConverter and accidentally scale it to 1200% instead of 1200 pixels – but be sure to have all your files saved or a lot of time at your hands before doing that). Afterwards, no part of the system will be very responsive and it takes very long for things to recover – if they recover at all.
- Shutdown
- In principle it could be argued that we are living in the times of ‘always on’ machines and we don’t need to shut them down. In particular as most machines these day’s don’t really turn themselves off properly but remain in some standby state that still consumes energy. And that idea is working beautifully for the iPod, say. But many people don’t like that exact feature of the iPod. I guess that is because being able to turn a machine off gives you the safe feeling of being in control and having the last word. That’s important of course.
Thus I conclude that – ideally – we wouldn’t need any of those features. Because our computers will reliably switch to ecologically reasonable modes when they aren’t in use. Modes in which they preserve our data and from which they recover instantaneously. In addition, the computers would reliably know when they can go to such a sleep mode and when they cannot. They won’t be fooled into remaning active by some hyperactive applications just as they won’t turn off ssh access to themselves just because you didn’t press any keys for a while.
But today’s computer hardware and software is nowhere close to that ideal world.
What confused me a bit about the post on the OS X sleep/restart/shut down options was that he claims they were designed for OS X. But the old Mac OS had the very same dialogue box that appeared when hitting the power button and everything worked in the same way.
In fact, things used to work much better in the old Mac OS. Because with all its flaws it had one fantastic feature: It was very predictable about how it would react to your keypresses. Even if the machine locked up for a short while, it would remember the keys you pressed in the meantime and run them as soon as it could.
OS X often fails when it comes to this: Not only is its window management somewhat intrusive and you can easily end up finding yourself typing into the wrong window just because something ‘popped up’. But even worse, sometimes keys you press just seem to get lost. And of course OS X still locks up. Just that it isn’t called ‘locking up’ these days but the wonderful technically exciting term ‘swapping’ is used for the fact.
And when you are pressing your laptop’s power button or Command-Eject to get that shut down window you will probably be in a situation where the shut down window either has to be loaded for the first time or where it has to be swapped in again after many hours of not using it. This gives you a second or two before you actually see the shut down window. And hitting the ‘R’ key, say, (for ‘Restart’, quite a handy feature of that window – even though it’s a bit strange on a German system where Sleep is called Ruhezustand
and Restart is Neustart
) in that period of time will usually lead to it being ignored by the computer.
And that, of course, is seriously bad UI. Not a problem of the shut down window, I suspect, but a problem that I frequently notice in OS X. A problem that Apple hasn’t solved in years. And a problem which I don’t think they are good enough to solve. Yet it is a big usability problem of OS X, because it means that from time to time you cannot blindly operate the computer, just knowing that it will process all your key presses predictably. Rather you have to keep an eye on the machine’s current state and possibly delay your actions a bit until the machine has caught up. The human adapting to the machine – rather than the other way round.
