It is kneecapped.
He invented a thing and left out a well known and understood core function required in the problem space, deliberately, not through oversight.
That's what makes it kneecapping.
The whole useful thing, ie the idea of a data format, including both knees (ie all the basic features any such thing needs) was there. The concept and necessity of annotation was a well known thing by then, and indeed he knew of it himself too, said so himself, and actively removed one functioning knee from that whole, to produce only the kneecapped thing.
He defined a kneecapped thing. Or he kneecapped the design. Whichever way you want to say it.
The difference would be if it was 40 years earlier and you are taking the fist stab at designing any kind of data format and comments just never occurred to you yet.
This is more like making a new programming language and deciding that it shall not have one of the basic math operators. If you need to multiply, tough shit, do it some other way. Just pipe it through JSMulti.
But he designed it with intention, as stated. It wasn’t like “ha ha suckers, no comments for you!” He just didn’t think it was a good idea and had a reason for it.
Are inventors required to only innovate a little bit? He released it, and there was every opportunity for someone else to offer something better and achieve the same level of uptake.