Rise and fall times, that is.  I have finally been able to measure the rise and fall time of these fast little single-gate packages. I was previously limited by the 1.2ns rise time of my old scope, but I suspected they were faster than that. Now, I have proof. The rise time is around 650ps and the fall 460ps. That's pretty fast.

This gate is being run at 3.3V with a 47-ohm series termination resistor through a short cable into a 50-ohm terminated scope channel. The waveform looks pretty good.

I have some 74AUC1G04 gates ordered, which have a shorter propagation delay, but I'm not sure about the edges. They supposedly use some kind of 3-mode output stage that dynamically changes its impedance to drive lines accurately. Or something like that; the datasheet isn't very specific.