HCI is the study of how people interact with computing systems, and the design of systems that fit human needs and abilities rather than forcing people to adapt to the machine. It draws on Computer Science to build the system, Cognitive Psychology to understand how people perceive, learn and make errors, Design for the visual and interactive layer and Social Science to understand how people work together. It's fundamentally an interdisciplinary field rather than a subfield of Computer Science alone.
Useful means a system provides functionality that actually addresses a real need or task. Usable means people can operate that functionality easily and without confusion. The two are independent from each other: a system can be packed with the right features but be painful to operate, or it can be smooth and pleasant to use while not doing anything a particular person actually needs.
Example: a universal TV remote with 50 buttons, covering every possible function, is useful because it genuinely can do everything someone might want from a TV. But for my parents who just want to turn the TV on and change the channel, it is not usable, since they can't find the two buttons they actually need among the other 48, and the small, unlabeled buttons make it worse.
A common example is a door with a flat panel handle on both sides, where you can't tell just by looking whether to push or pull. The person's goal is simply to walk through the door, but the hardware gives no signal about which action is correct, so people find out only by making the wrong move first. This is widely known in design circles as a "Norman door", after Don Norman, who used it to argue that good design shouldn't need printed instructions to be understood.
Improvement: use a flat plate only on the side that should be pushed and a handle or bar only on the side that should be pulled, so the shape of the hardware itself signals the correct action without needing a label.
A prototype is a rough, early version of a design, such as sketches, a clickable mockup or a physical model, built to test ideas before investing in full implementation. Testing a prototype with real users surfaces problems a designer can't see from the inside, like where people hesitate, misread a label, click the wrong element or give up entirely.
Asking "Do you like it?" wouldn't tell you whether people can actually use it. People are often polite, and liking how something looks is a different thing from being able to complete a task with it quickly and without errors, so usability testing needs to watch behavior rather than just collect opinions.
KIBO is a screen-free robotics kit for children ages 4 to 7. Kids program it using physical wooden blocks that snap together, with no typing and no screen involved at all. It benefits young children who can't yet read or type fluently, letting them learn sequencing and computational thinking through tangible, hands-on interaction instead of text-based code.
To test whether it helps them, you wouldn't rely on a young child's self-report. You'd observe instead: can they build a working sequence, do they debug it themselves when the robot doesn't behave as expected, do they stay engaged and does learning hold up across repeated sessions.