In the 1950's, any computation done on a computer was performed by programming in machine language, which was tedious and hard to follow (if you've ever seen an assembly language code, imagine it without the mnemonic, and you would have a good idea what it was like).
This was actually slowing down the development of code, and John Backus proposed that IBM allowed him to set up a team to search for an alternate approach, that of developing a program that would translate high level, readable instructions into assembly. The result of this initiative was the development of the first Fortran compiler, which proved good enough -- the resulting assembly was almost as efficient as if it had been directly coded by an expert using all the tricks in the book.
One note here, at the intention of anther answerer: the GOTO is NOT forbidden. Some do claim that is an ugly construct, and could lead to confusing code, which may be a point (well written, a code could be as decent using GOTO as using any other sturtured appraoch). But the bottom line is, at the machine level, in assembly/machine language, there are ONLY GOTO's (no WHILE, no REPEAT, no ELSE; all those instructions unroll to a bunch of GOTO's once compiled), so even the recommendation to avoid GOTO (the famous "GOTO considered harmful" by Edsger Dijkstra, which prompted a rebuttal called "'GOTO considered harmful'considered harmful", which escalated to at least two more layer of "considered harmful" encapsulation) cannot be taken as an absolute.