Operating System Services, Application Software and Character Encoding

What the operating system does, and for whom. Applications never touch hardware directly; every file read, every pixel drawn, every packet sent goes through the operating system. That is why it carries a fixed list of responsibilities you should be able to name and explain: memory management (allocating RAM to programs and keeping them out of each other's space); processor scheduling (deciding which process runs next); the file system (organising storage into files and folders with permissions); device management (drivers that translate generic requests into commands for a specific printer or disk); the user interface; security (users, passwords, access rights); and multitasking (switching between programs fast enough to appear simultaneous). A question that asks for a function wants the function and what would go wrong without it.

Hiding complexity. The most examinable idea about operating systems is that they present a simple picture of a complicated machine. A physical disk with platters and sectors appears as a drive letter and a folder tree. RAM plus a region of disk appear as one large, continuous virtual memory. Thousands of printer models appear as one Print command, because the driver does the translation. A Java program sees a single virtual machine rather than the real processor, so the same code runs on any hardware. Each abstraction lets application programmers and users ignore detail they do not need — at the cost of some performance and of not knowing what is really happening when something goes wrong.

Application software. Word processors, spreadsheets, databases, browsers, email, graphics and computer-aided design each solve a class of user problems and share common features: a graphical interface, toolbars, menus, dialogue boxes, and — the point examiners like — common features that reduce learning, so that Save or Undo works the same way in every program. Applications are chosen by fitness for purpose, cost, compatibility with existing data and hardware, and support, not by feature count.

Characters are numbers. A computer stores only bits, so every character is a code. ASCII uses 7 bits for 128 characters: uppercase A is 65, lowercase a is 97, the digit 0 is 48, space is 32. Uppercase and lowercase differ by exactly 32, one bit, which is why case conversion is a single bit flip. ASCII has no Vietnamese, no Chinese, no accents; Unicode assigns over 140 000 codes covering every script, and UTF-8 stores them in one to four bytes, so Vietnamese text takes more space than English. Because digits are characters with their own codes, comparing numbers stored as strings goes wrong: "10" sorts before "9" because the character 1 is less than the character 9. Hexadecimal is a shorthand for binary — each hex digit is exactly four bits, so 1010 1111 is AF — used for memory addresses, colours and MAC addresses because it is compact and converts trivially. The machine does not store hex; hex is for human eyes. Underlying all of this: the same bit pattern is a number, a character, a colour or an instruction depending only on how the program reading it chooses to interpret it.