If I was to explain LLVM in a painfully simple manner, I might say something like: “LLVM is a framework that allows you to define a syntax for interacting with computers, and instantly have it work for all computers”. This description does a colossal disservice to the scope of the project, but it serves to illustrate the basic idea: there are many languages (frontends) and many CPU architectures (backends) and we don’t want to re-implement every translation permutation.
