Jargon is inevitable in any sufficiently advanced field of endeavor. Here are some terms frequently encountered when working with decompilation.
A human-readable rendering of machine code. Assembly language consists of statements, which in turn typically consist of an optional label, an operation or pseudo-operation, and optional arguments:
; label operation operand(s)
program_start: push ebp
mov ebp,esp
A binary file, or simply binary is the input that Reko decompiles. These typically consist of machine code. Many binary file format contain additional metadata which assists the decompilation process by providing more specific information about the contents of a file. For instance, such metadata can specify that some sections of a binary are executable while others aren't; symbol information can specify the precise locations of objects the decompiler wants to discover.
Intermediate representation (IR) is a machine-indenpendent representation of executable code. Reko converts machine code from different architectures into this representation so that it can use common algorithms during the decompilation process.
Machine code consists of instructions directly executable by a computer. The instructions are encoded in an architecture specific way. On some architectures, every instruction is encoded in units of the same size (e.g. bytes, 16-bit words, 32-words); on other architectures, instructions can be encoded using sequences varying numbers of these units.
Rewriting (called 'lifting' elsewhere) is the process of converting machine code to intermediate representation.