Tablets fail because the stylus is a mouse simulator

Microsoft’s ultra-mobile format is about a year old and it has not worked. The machines are expensive, they run hot, the battery life is poor, and the ones I have handled feel like a laptop that has been folded up rather than a thing designed to be held.

The diagnosis being offered is that the hardware is not ready. Give it two more years of processors getting cooler and batteries getting better and the category will arrive.

I do not think the hardware is the problem, and I think that diagnosis is why the next attempt will fail too.

Every tablet built in the last five years has been made the same way. Take a laptop, remove the keyboard, and ship it with a desktop operating system. And a desktop operating system assumes things about how you point at it that a finger cannot provide.

It assumes pixel precision, because the close button on a window is about sixteen pixels square and a fingertip covers considerably more than that. It assumes a hover state, because menus open on rollover and tooltips appear when you rest without clicking, and a finger has no way to be near something without touching it. It assumes a right button. None of that survives contact with a hand.

So manufacturers supply a stylus, and the stylus restores the precision, and it also means the device only works when you are holding a small plastic stick that will be lost within a fortnight. A stylus is not an input method. It is a mouse simulator, present so that software written for a mouse can keep running unmodified.

The demonstration Apple gave in January is the interesting counter-evidence here, and it is not interesting for the reason most coverage says.

The multiple touch points are a nice trick. The significant part is that everything on that screen was drawn at a size a finger can hit. The targets are large, the lists scroll with momentum rather than a bar you must grab, and there is no hover state anywhere because the interface was designed knowing there could not be one. Somebody wrote the whole thing from scratch for a hand instead of adapting something built for a pointer.

That is the expensive part, and it is why the PC industry cannot do it. Dell and Toshiba and Samsung do not write operating systems. They buy one, and the one available assumes a mouse, and no volume of cheaper silicon changes that.

Which means the tablet does not arrive when the components get good enough. It arrives when a company that controls its own software decides to build one, and there is a short list of companies in that position.