Perhaps designed to do too much, this wall-mounted interface was customized to control all the audio/visual components, telephony, and lighting for a large conference room. Attempting all-in-one simplicity, the device instead presents its users hodge-podge complexity. Two of its issues – and some solutions – are explored below.
Because the hardware provides a number of controls that are partially duplicated on the custom program the device runs, the user starts in Norman’s ‘gulf of execution’, without clear
affordances and proper constraints to guide her intent to appropriate action. If a user wants to go to the home screen, she can hit the house-shaped icon on the device or the ‘Home’ button in the program’s top navigation. Similarly, to power off the device, she can use the ‘Power’ button on the device or the one in the navigation on the screen.
This duplication – a missing logical constraint – unnecessarily forces the user to choose between two buttons for the same function. After integrating this duplication into her mental model, perhaps the user then decides to change the lighting controls and tries the light-shaped icon on the device similar to the one present in the top navigation. This time, nothing happens, with no feedback of any kind!
So in addition to poorly constraining users with duplication, it’s not even consistent here and fails to provide feedback after the system fails to adhere to the faulty conceptual model.
The inherent issue is that the hardware contains a set of buttons that either a) constrains the the software designed for it or, b) renders the software that does not abide by its limitations inconsistent and confusing. If the hardware maker intends for most users to develop custom software for their product, it is best for the device to include only the ‘Power’ button and to leave the rest to the program’s software.
As for the program designer working with what now exists, they should not attempt integration with the hardware controls unless the mappings between these buttons and their functions are clear, unique, and meaningfully distinct from the kinds of functions available in the program. To fix the current situation, I’d recommend removing the ‘Power’ button from the program’s navigation menu and physically covering or otherwise hiding all hardware controls but the ‘Power’ button. While some methods of covering these controls may appear blunt and unattractive, in this struggle between aesthetics and usability, usability should prevail.
ISSUE: LIGHT SETTINGS
Being able to easily and intuitively control both natural and artificial light in a large conference room that has a wall of windows and multiple light fixtures is magical. This, in theory, is what the touch screen’s light settings screen offers, but let’s see what a user would actually encounter using this Crestron device.
The settings include a section for light and window shade presets, a section for light settings, and a section for setting the position of the window shades.
Although these lights and shades correspond to visible objects in the real world, the layout of this screen provides no useful mapping to this reality. Not only are the buttons ordered in a rather rote and thoughtless way, but the screen affords no evaluation of the system’s current state. Whether the room is pitch black with lights off and shades down or blinding bright with all lights on and shades up, the display appears the same.
Of course, in either of these extreme cases, the user could gather knowledge from the world to know the system’s state, but many more cases would not be clear at all.
To improve upon this design, the controls for lighting must map better to the reality they act upon and make more visible the current state.
Setting the position of the room’s shades could map more naturally to the manual behavior of adjusting a shade — a slider that moves up and down. Similarly, position the light-related buttons as they are positioned in the room. Group them as being near the front or rear of the room, with upward facing lights on top of downward facing lights. Add to this some visual indicators of the current status of each light and both visibility and mapping quality are improved.