I disagree. It should at least cover what aspects should be considered when designing an algorithm for use in the real world. You dismiss the entire field of experimental algorithmics as if it were nothing more than a bunch of algorithmicists who know some computer architecture things, but there’s a lot more to it.

I’m not arguing against teaching classical algorithm design and analysis, they’re both incredibly important imho. But we should stop ignoring the shortcomings of the models used, hand-waving them away with “the rest is just a bit of engineering”. It’s not.