"I absolutely think the slowdown in computing performance gains play a big factor [in slowing PC sales]. Maybe even more so than the whole tablet thing. Why would you replace your PC if it’s not noticeably faster than what you bought two or three years ago?"
see Why Moore's Law, not mobility, is killing the PC
Of course that’s a factor but there are others:
- Moore’s Law also helps make smaller, cheaper computers without fans and storage drives. Without all those moving parts thin clients and tiny desktops will easily do on the tenth year what they did on the first year. You can upgrade servers or add servers much more cheaply than you can roll out a new fleet of client machines. It just is too costly to change all the hardware, something that prompted a lot of migrations to GNU/Linux because it required fewer resources to work.
- The typical CPU in a single user system idles most of the time and the real delays are accessing dozens of files to open the next window or application. One can update storage to something faster without changing the rest of the machine but one can also do the heavy lifting on a server that can have enough RAM that storage doesn’t need to be accessed for applications. They can all be stored in RAM and shared amongst users. Zero storage is obviously faster than any other form of storage. Again, thin clients and small cheap computers work.
- The ARM processors are clearly able to do the job most users need done and consumers see the advantages of a pocket-sized PC over the legacy boxes and even notebooks. Hundreds of millions have traded off the larger screen for the greater portability. This is partly an effect of Moore’s Law making ARMed machines good enough but it’s an entirely different architecture with its own advantages. If there were no ARM, salesmen probably still could sell the next release of Wintel…
- Today’s networks are a huge improvement on even five years ago. Gigabit/s is fairly standard on the LAN and Internet connections for individuals approach the pipe whole buildings used to have to the Internet. This means a thin client and its server are behaving more like one machine and the user appreciates the web application that runs on someone else’s hardware with all those headaches gone rather than having the headache over Wintel at home. No headaches. No need to replace the PC to fix the headaches, temporarily. Wintel was built on making new sales to soothe headaches created by Wintel. That worked when Wintel was the only choice on retail shelves… Welcome to the 21st century.
- It’s also a matter of software. Peoples’ small cheap computers are not slowing down because they have less malware and incompatible and resource-hungry apps on the go. They are consuming less power, making less noise and taking up less space. They are often far more portable than Wintel to boot. The app-stores and package managers of */Linux system just work better for people customizing/managing their personal computers. Clearly, people like installing an app in seconds without having to fiddle with re-re-reboots, malware and Patch Tuesdays. Those negatives are all associated with the legacy PC so why buy more headaches? Wintel might have adopted and extended ARMed systems a few years ago and rationalized their system management and headed the stampede off at the pass but they didn’t and we are seeing a quite different historical change.
All in all, this is a good thing about which I have been writing for years now, first about GNU/Linux instead of that other OS, then APT package manager of Debian GNU/Linux and then small cheap ARMed computers. It’s empowered by more than Moore’s Law, but it is empowered and Wintel is getting taken down a level or two. It’s no longer essential and soon Wintel will have to work for a living competing for price/performance on retail shelves. I expect 2013 will be remembered as the year small cheap computers really went on a tear.