I'd like to see the ALGOL group writing "we're faster and leaner than fortran" while writing their spec and no compiler to run tests against. I struggle to follow.
Something like this actually happened when they had to decide between Wirth's and Hoare's practice-oriented proposal and Wijngaarden's rather academic proposal. There was quite a bit of controversy about this, not based on true implementations. Big egos exist everywhere and in all times.