Must Everything Look The Same?
Way back in 1983, I put together a team to design a real-time screen-based information display system. This was ten years before the good folks at Wired sat down to design one of the very first commercial (as opposed to academic) web sites. What the two teams, ten years apart, had in common is that neither of us really had much of a clue how to go about our tasks. There weren't many models around so we just made it up as we went along.
The system my team was replacing consisted of four TV cameras pointed at boards. Data entry consisted of writing things on the boards, page selection consisted of changing TV channels on the display terminal. When we came to design a computerised version, it turned out that there were all sorts of things for which there was no model - what would now be called "site navigation", for example.
Although our users would receive some training, their main attention was on another task entirely, so the system had to be very intuitive and keystrokes had to be kept to a minimum (Although I think the concept of a "mouse" existed, they were in labs and none of us would actually see one for another few years).
Moore's law applies, so the system supported twenty-five display pages containing about the same level of content as now pits on a palm-style PDA screen. The twin central servers occupied a six foot high nineteen inch rack, but twenty years on, bits of that system design are still in use, day in and day out, so we must have done some things right.
I've since designed two more systems for the same purpose and kept many of the core elements, getting a few more things right each time and taking advantage of some, though by no means all, advances in technology.
It turns out that there are some pretty hard and fast rules about user navigation. Probably not as many as Jakob Nielsen would have us believe, but also probably more than most designers want. Does that mean that all web sites that do the same thing should look the same? I don't think so.