Well, okay.... I'm merely playing devil's advocate here, but I can see the value in leaving more and more of the day-to-day avoidance and decision-making in our kids' hands at this stage.
I base what I ask for from the school on what we're (and our physician is) comfortable with in terms of day to day management. That is, we do allow (encourage?) our DD13 to go to the movies, visit local coffeeshops, etc. without us in tow, and WITHOUT controlling the food FOR her...
Like Mac's son, this has been miraculous for her socially. I can't possibly overstate it, truly. On the other hand, it's more than occasionally terrifying for us as parents. But so far, she's managed very very well. She is pretty much never without one or more friends that knows about her allergies-- and several of those friends have asthma.
As long as autonomy is being HANDED OVER to the allergic adolescent, then I have less problem making the 504 plan more "real-world-like."
Where I have a problem with this in a school setting is, as noted above, when the student is FORCED to remain in a situation which is: a) uncomfortable, and/or b) overtly dangerous. We hand over control and let DD venture places that we'd NEVER have taken her four years ago-- but the reason is that SHE is in control of herself and she has carte blanche for managing. She can (and has) gotten up and left, or opted out completely, when a situation seems unsafe to her.
I also have a BIG problem with food rewards-- because for our kids, that directly inverts the external motivator and makes it punishment for doing a task/activity. PUNISHMENT. That's actively damaging to kids that see it that way. If it isn't, then there's no point in offering it as a 'reward' to any kids, since it's not really a "reward." KWIM? If it is a 'reward' for those who get it, then it's PUNISHMENT for kids who find it a hazard rather than a pleasant thing. That is just such simple operant conditioning that it boggles my mind that ANY educator could possibly fail to note it.
Secondly, extrinsic motivators have been shown, in adolescents and adults, to be PARADOXICAL. That is, tasks associated with rewards become LESS intrinsically rewarding, LESS appealing, and less competently accomplished. There's research on this effect, by the way. I'd use it. For kids with developmental delays, fine. For adolescents without them-- not so fine.