[cess11]

I kind of agree with both of you. I would have become better at Prolog faster if I had stuck to the discipline Triska promotes early on, but I also think the guidelines you cite are reasonable once one is starting to do somewhat productive work where compromises and speed factors in.

Over the years I've come to a similar position in other languages as well. If a functional-ish solution fits performance constraints and is maintainable, don't mutate or reach for global state, things like that.