The people who argue dogmatically against any use of gotos are missing Dijkstra's main point. While the title of the paper is about gotos, the body is mostly concerned with discussing how programs can be structured so that it is feasible to reason about their correctness.
When this paper was written, gotos facilitated the rampant production of confused and confusing spaghetti code, and many programmers believed gotos were required to write certain types of program. At that time, banning gotos seemed like the only thing that could fix this. Since then, we have learned a couple of things:
1) Good programmers will write well-structured code even when they have the option to use gotos (and even if they actually do.)
2) Confused programmers will write confused code even while adhering to structured syntax rules.