A shift-reduce parser is a program that is generated from a grammar definition. It is a kind of state machine capable of doing two kinds of actions:
Either it consumes a token from an input program, adding up a new state in a stack. This is called shifting.
It recognizes that a rule of grammar has been fully matched, so it pops as many states from the stack as the rule contains and acknowledges that it recognized that particular rule, adding up an entry to the parse tree. This is known as reducing.
Given how these parsers operate, we'd say that they belong to the "bottom-up" parsing family. That is, they operate on input and deduce the parse outcome while traversing it, in contrast to the other way around, which is starting from the rules of grammar and finding structures in the program that obey them. Our parser will then operate on a linked list structure, consuming tokens one after another.
Now, a particular state of the parser depicts the situation it...