Emergency doors (not "exterior doors" as a whole) are to open out because there might be a barrage on them in an emergency. Your house door isn't likely to experience a barrage (or rather, the fire code forbids you from having too many people in your house, partly because it isn't fitted with proper emergency exits for crowds).
This has nothing, what so ever, to do with whether or not a door is intuitively designed. Also, whether or not you lose a second opening a door is irrelevant (and you've just given the excuse for 1990's enterprise software UX - as long as it's possible to complete a task, who cares if it's easy or intuitive. Everyone, it turns out, but it took a while to hammer that point int).
The point is that "we" have the knowledge to design a door that is clearly and immediately usable (without losing a second) by anyone who's ever used a door - but sometimes "we" apparently chose not to use it. That is the conundrum outlined.
I know it isn't exterior doors as a whole. But the vast majority of doors you will encounter on a public building have to be usable in an emergency. So, I think you'll be hard pressed to find one that doesn't open out. (Again, unless you drop into residential doors.)
Your point that we have the knowledge to make doors that are immediately usable is just something I don't believe. I've seen people use doors incorrectly that fit the guidelines of that book. Hell, I think I've done so myself. And, I argue that this is almost completely related to the consistency of how a door opens. Not the handle. (e.g., after years of living in a high rise where the entrance opened out and all I did was go to public buildings where all doors open out, not too shockingly, I was more prone to opening house doors incorrectly.)
My argument was never that "so long as it is possible..." The argument I put forth was more that the cost for getting a door wrong is ridiculously low, so people don't put that much thought into it. On either side of the isle. (Well, designers want doors to be pretty.) Your 90s example is attacking things where the cost was actually high, but the designers didn't care.
I think you can basically sum up my viewpoint with the common attack on intuitive designs that nothing is truly intuitive in and of itself. Pretty much everything is leveraged on something you have previously learned. If it seems intuitive, it is really just familiar. In the case of doors, the familiar behaviour trumps any design decisions. (If you have some good studies against this, by all means I'm game to read them. More than just the book that started this thread, though.)
This has nothing, what so ever, to do with whether or not a door is intuitively designed. Also, whether or not you lose a second opening a door is irrelevant (and you've just given the excuse for 1990's enterprise software UX - as long as it's possible to complete a task, who cares if it's easy or intuitive. Everyone, it turns out, but it took a while to hammer that point int).
The point is that "we" have the knowledge to design a door that is clearly and immediately usable (without losing a second) by anyone who's ever used a door - but sometimes "we" apparently chose not to use it. That is the conundrum outlined.