here's my thoughts:
First off, don't consider interfaces just with computers. We deal with user interfaces all the time - cooking with a stove/oven, doing laundry, adjusting the thermostats, playing a video game, using a car wash, making a telephone call, watching TV, etc. In fact, most people interface with other things far more than they do a computer on an everyday basis.
Second, interfaces are going to be specialized to the task at hand, as well as personal preferences. A given terminal station may be usable for a variety of tasks, but one configured for navigation/piloting is going to look a lot different than one configured for gunnery or engineering. Given some of the specialization needed, it's highly likely that not every terminal will be configurable for every function, either. Certain types of functions may need specialized terminals - I hardly expect the entertainment terminal in the ship's rec room to be able to handle piloting or engineering functions.
Third, in most cases, the interfaces only need to be as complicated as the task at hand, which means that many functions are more likely to be performed by dedicated interfaces. There is no point to installing a configurable interface terminal to control a door, a washing machine or an oven, for instance, because no matter how fancy they get, they have limited functions. Even if the configurable terminal is cheap enough to warrant using one instead of building a specialized interface, more than likely the interface will be locked for that single function.
Some other thoughts on specifics
- Voice: voice is only practical in environments where only 1-2 people are going to be speaking. I think it will be reserved more for the interfaces in private rooms/offices.
- tactile feedback increases a person's ability to multitask. Think about that receptionist in the office who can be handling incoming phone calls while typing their bosses handwritten scribbles into something coherent. Or the ability to dial a phone in the dark. Holographic controls and all are nice, but if there is no tactile feedback, a pilot is going to have to pay attention to where his hands are, instead of the display showing where the hostiles are.
- certain tasks lend themselves better to a physical control, oftentimes a specific type of physical control, regardless of what technology is capable of. With drive by wire, we certainly have the capability to control a car using a touch pad, but anyone who's played any type of racing game on their xbox or playstation knows it's much harder than turning a steering wheel.
- security. You already mentioned biometrics, which is probably the way the future will move. But any good security system will involve 2 or more methods in combination, and physical security methods like access cards/keys/biometrics etc are inherently more secure than things like passwords and user ids, but in combination they are orders of magnitude more secure than either alone. The real problem with invisible security is that it might make those who bypass it "invisible" as well.
- failure. As a technician, I always ask myself - "What environment is this going to be used in?" and "what happens when it breaks?" The harsher the environment, the more likely it is to break, which means the more the more reliable the technology has to be. Furthermore, the technology chosen might need to work under less than ideal conditions. Take the fancy Iris valves used by most Traveller ships - they're nice fancy high tech powered alternatives to manual hatches, but IMTU, they all have a manual hand driven back up to open or close them during emergency situations.
Finally, the real thing to keep in mind - Traveller pretty much presumes that humans haven't changed much in the several thousand years we've been out in space. So no matter how fancy the technology gets, unless it's a neural mind-reading interface, humans are still relying on their five senses, and using the same extremities we always have to manipulate machines. Certain things will change - frex, it is faster and more efficient to push buttons on a phone than it is to use a rotary dial. Configurable touchscreens are more versatile than a standard keyboard in remapping their functions, but in the end if you would find it ackward to use now, then it will probably still be ackward to use in a few thousand years.
In particular, in the mid to late 1990s, I worked on a search engine deployed in public libraries so the public could find out what was available. The main thing I learnt from that is that it takes care and dedication to produce friendly user-interfaces.