Hi Eric,
Yes i believe that too, but then again all the spec's i've ever seen dont specify a time or an energy level, just a voltage level such as 5v, 10v, etc. In other words, they will spec an input pin as allowing some maximum voltage level like Vcc+0.5v or something like that...ie they dont say "Vcc+0.5v for 1ms, Vcc+1.0v for 100us, Vcc+5v for 10us", they just say "Vcc+0.5v" and that's it. This is the way it is in the transistor world too, where the breakdown is considered to take no time at all once the threshold has been exceeded. This is probably why they never specify the time period or the energy unless it has some other significance.
Also, it would be hard for us to access the damage with varying voltage overshoots and time periods. For example, we might know 10v for 1000 seconds is damaging, but how do we access 10v for 1us for example...does it do damage or not? The only way i myself can be sure is to never allow the voltage to exceed the spec sheet max spec cause that's all the concrete knowledge i have about that kind of specification.
This is pretty much standard practice though, with inductors too as im sure you know. With capacitors in some pulse circuits there's also the chance of the voltage going too negative, in which case a diode clamp to ground is used. It's more common with inductors, but caps can also be a problem.
Make sense?
Another problem that can come up is with some input driving the cap and a clamp diode to Vcc. The pulse energy is high enough and the total impedance on Vcc is high enough such that the pulse can drive Vcc higher (surprise surprise ha ha). This puts a high pulse on the Vcc line and of course blows out components that cant stand an increase in Vcc