"Well, compared to other steels it is extremely strong yet more brittle."
You're right. You hit the nail on the head right. But it is strong. And when your cutting edge fails, it was too weak for the force you were using. Some steels at high hardness exhibit a brittle failure, meaning no distinct yield point, but a break. D2, at a nice blade hardness of 60-62 will yield. And the edge doesn't chip off (like some hard S30V, a "tough" stainless) it rolls. This is for a "cutter", not a "basher".
D2 is more brittle than other common shop metals like A2, S7, H13 or P20 by the joules of energy dissipated in impact tests. But it has been my observation (as a machinist, not a knife maker) that at a given hardness, D2 is stronger, and thus exhibits less failure as a cutting tool, brittle or otherwise. It gets used as blades to grind rock contaminated plastic, not A2. I imagine that in a high impact application like a sword or cutlass it would not be tough enough.
BTW, those impact tests in the literature were for the "standard" heat treat, which is not the best heat treat for a knife. The volume of large carbides does not help the toughness. Nor the sharpenability. I think a fast quench makes for a better knife blade.
"The simple fact someone makes blades or punches that hold up well doesn't mean the steel is not brittle, because brittleness doesn't automatically mean the tool cracks under normal or even hard use. It just means compared to other, comparable steel types it is more brittle, due to it's structure."
I agree with that. But that doesn't mean it deserves a reputation as a brittle blade, because I have observed good durability with D2 at high hardness (63-65). And lousy durability with O1 at the same hardness.
So let me modify my statement. For a moderately high hardness application, D2 has good durability and isn't too brittle. And a fast quenched blade with a fairly thin grind isn't a bear to sharpen and is a joy to own.
