a file access method

Suppose you have a long playlist. You could listen to it straight through from the first song to the last; you could jump directly to song number 27; or you could keep an index of artists so you can leap to the first song by a chosen band. Those three habits are the three classic file access methods: how a program chooses to move through a file's contents.

Sequential access reads or writes the file in order, one chunk after another, with the position pointer simply marching forward. It is the simplest and matches how we read text or play audio; rewinding to the start is allowed, but you generally go front to back. Direct access, also called relative or random access, lets you jump to any record or byte offset immediately, read or write there, then jump somewhere else, with no need to pass through everything in between. This suits a database that must fetch record 5000 without reading the first 4999. Indexed access builds a smaller index file that maps a key (say a customer ID) to the position of the matching record; you search the compact index first, then use direct access to leap straight to the data. Indexes are themselves often stored in files and accessed directly.

Why it matters: the access method you need shapes the file system features and even the hardware that serve you well. Sequential access is friendly to spinning disks and tapes (no jumping around); direct access is what makes random reads on databases practical and is cheap on an SSD where there is no seek penalty. An honest point: these are PATTERNS of use built on the same underlying read, write, and seek operations, sequential versus direct is mostly a question of whether you call seek and where, not a different set of system calls.