All problems got considered and found moot. Don't worry it will work :-) Even while it violates IrDA specs (which is a reason we don't usually see it being done, along with power consumption reasons) IrDA transceivers always come with LED and photodiode in a module, nobody usually puts a second IR LED in just for fun, after all it's IrDA and not meant for CIR.
I might be wrong but I thing Copernicus meant over-saturating our own receiver. We might need to reduce the output power in full duplex mode. Actually, maybe even in half duplex mode since saturated semiconductors tend to take some time to get out of the saturation.
IrDA data communications operate in half-duplex mode because while transmitting, a device’s receiver is blinded by the light of its own transmitter, and thus, full-duplex communication is not feasible. The two devices that communicate simulate full duplex communication by quickly turning the link around. The primary device controls the timing of the link, but both sides are bound to certain hard constraints and are encouraged to turn the link around as fast as possible.