The thumb pad interface,smart phones

SmartPhone have a one bit interface.

The user controls the tacking up util the thumb is at th screen, so position accuracy is determined by the icon crowding, and guess what, the icons are all the same size. The actual had and user need agree, the button is pressed or not.

The reason is that our hands are better off throwing he things, they feel like a rock an our thumbs are adapted to rocks. Thumbs do not get icons, never will. So, comfort level is again, very low rate that user can do it without concerted effort. In the Fitts-Shannon sense, the one size fits all is a one bit system.
The very first smart phone, invent by Sam in 1820, could su[ort maximum hand eye, in this case the fore fingers, at 20 five bit words per minute. Three bits a second was really good.  But that was temporal accuracy of the wrist, like throwing a rock, or making the three pointer; we was adapted for it.

The fore finder likely could do the three bits per second, laing down the smart phone, focusing on the task and having a high sample rate icon properly sized, make it a telegraph key. Linux folks know how to talk to telegraph keys.

It is sound theory, and something to keep in mind, like the default zoom up step, do not make it linear, make it take the step needed to put your complex application widgets into a 3 bit resolution for the professional human interface. So,k for example, at detail level, there are basically eight different size tpes, obvious,. That is a lot, for the average user, four widget type, small, and large button, text entry window and text blob, the text editor interface is four bits. The user can jump around accurately with four setting on his hand motion.

Measuring the consume bandwidth

The general method is give the consumer test group a smart credit cards, extract their balance sheet, he smart card will have already done the Huffman tree on the thing. The balance sheet is derived from an integration of the payment generator function,. If you could translate a 5 bit  uniform random thru a huffman tree, you generate the typical event values in a complete sequence. The balance sheet is an integration of the nodes, the nodes are a directed cash flow. So to understand the interface, look at the credit card transaction sequence.