Dijkstra's ideas may be relevant to safety-critical domains (at least to some extent) but the article is flagrantly ignoring cost-benefit tradeoffs. Empirically we see that (manual) proof-oriented programming remains a small niche while test-driven programming has been very successful.
He's certainly not ignoring cost-benefit tradeoffs. He acknowledges this as a perceived weak point, and claims that, when practiced properly, the tradeoff is illusory. (I rate this unlikely but possible, around 2% that it's purely true and another ~20% that the cost increase is greatly exaggerated.)
I'm pretty sure Dijkstra would argue (and I'm inclined to agree) that proof-oriented programming hasn't gotten a fair field test, since the field is taught in the test-driven paradigm and his proof-oriented teaching methods were never widely tried. There's defin...
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.