A few weeks ago my iMac (21.5-inch, late 2012, 2.7GHz Core i5, 8GB memory, then OS X 10.12.3) started seizing up, sometimes so badly that I had to do a hard reboot, other times just slowing a a crawl. I tried all kinds of things but could not afford the time to get systematic about pinning it down, I was desperate to churn out various memos etc.
Yesterday, I finally got more serious and found the main problem was that an Apple process called “suggestd” was taking up nearly 100% of the CPU time, presumably locked in some kind of endless loop. The forums suggested this problem first appeared a few years ago and was apparently sorted – after a while. Then it re-appeared under v10.12.3. The second suggested fix I tried worked.
But removing that bottleneck revealed another. Looking at the log files I traced that to Snapz. The utilities that Apple provided used to be fairly hopeless for taking an image of an area of the screen or recording a video, so like many people I purchased Snapz. It proved useful for several years. I had to uninstall it when I moved to El Capitan (OS X v10.11), but I discovered that Apple had caught up and I no longer needed it.
However, it turned out that it was much trickier to uninstall than I had realized and I had not completely got rid of it. That bit me when I moved to Sierra (OS X v10.12). It turned out that launchd was trying to launch a LaunchAgent for it every 10 seconds, mostly this just slowed things down slightly but sometimes it glitched iTunes or froze for a minute if the TimeMachine backup was starting. I like TimeMachine, but there is not enough control over it to stop the wretched thing starting an unnecessary backup just as you have unsleeped the machine.
I cannot really fault Apple for the Snapz mess which was primarily due to poor uninstall instructions from its authors, but sorting it out is not made any easier by the lack of convenient Apple documentation for the key Apple process launchctl. I reckon to be reasonably competent at software – I know in broad terms what to do, but I am not that fast at it, because I try to limit my time on playing software to a few per cent of my time. To get really good you have to do it full-time, then you remember the thousands of tiny details and do not need to look them up. Be that as it may, the whole mess (suggestd and Snapz) has wasted many hours of my time at a particularly busy period. Grrrrr!
I never quite know the answer to the complexity of modern desktops. Most people only want fairly limited functionality, and delivering that could be really cheap – £50-100 for a smartphone or tablet and £100-200 for a laptop or desktop. But that does not suit Apple, Intel or Microsoft.
The hardware guys are stuck with Moore’s Law, they every year or two the capability of hardware improves by a factor two. So in principle they could halve their prices and deliver the same – which obviously has little appeal to them – or they can deliver more for the same price. The “more” sometimes involves new things that are genuinely useful to the user: displaying good quality video; good touch screens. Indeed, after decades of failure, good voice control is finally starting to emerge (using extensive server software via the internet). But much of the time the “more” is of limited use or a positive nuisance.
What about the massive, bloated operating systems? To be fair to Apple, it realised that we had to move away from the idea of a single machine the PC, albeit sometimes clothed as a desktop and sometimes as a laptop. It introduced the smartphone and later the tablet. It is now struggling, via Swift, to tie iOS and OS X together better.
But it remains true that most people would be happy with much simpler cheaper desktop machines with a small set of standard apps and no ability to install new apps off the internet. Unfortunately no one has a financial interest in providing that, and it has proved possible to sway consumers with advertising focussing on fashion and snake-oil rather than functionality, reliability and ease-of-use.