Assembly Language

Computers are instructed to do what they do by bit patterns. These bit patterns can be represented in different ways. As positions of a row of switches, as patterns of 0 and 1 characters, as decimal or hexadecimal figures of the values to which these bit patterns would compute or as strings from a catalog where the "meaning" of the bit pattern would be expressed in an abbriviation of, say, 3 characters.

For example "LDA" would represent the bit pattern that instructs a certain type of processor to LoaD its Accumulator.

Assembly language then consists of a catalog of such strings representing each legal instruction pattern, and a syntax (mainly for adressing). Assembly language mostly also offers a scheme to go to a sequence of these instructions and back in a symbolic rather than numerical way (library call), patterns for symbolic memory adressing (variables) and hiding of rudimentary memory management (e.g. automatic allocation and deallocation of variable storage space).

It is easier for a human to use than the other mentioned representations (which also need translation). It is also easy to program a (nearly 1:1) "translator" for stored files of these sequences. These translation/library/memory management programms are what we call Assemblers. Many people use the word also for the designation of assembly language.

Humans can use their intimate knowledge of subleties in specific computer architecture to optimize critical sequences in these programs.

A.L. is different for each kind of computer or CPU. Only the innermost cores of operating systems and the most time critical parts of other systems are thus written, adapted, optimized and translated for specific machines.