Hello again,
The convention for rise time is 10 percent to 90 percent, but that is when no other facts about the application are given. As i pointed out, for calculating the slew of a perfect ramp we can use just about any two percentages, as long as we measure the time between those two measurements as well.
The slew rate is:
sr=rise/time
where
the rise is the change in voltage, and the time is the time that change took. So if the lower voltage is 1v and the higher voltage is 2v and it took 1us to rise from 1v to 2v, then the slew rate is equal to the rise (change in voltage) which is 1v, divided by the time which is 1us, so we get 1v/us as the slew rate.
10 percent to 90 percent is often used because often the signal behaves abnormally between 0 to 10 percent and also between 90 and 100 percent, and those levels are often not significant to the effect it has on the circuit so we dont use anything less than 10 percent or greater than 90 percent.
In other words, the circuit would produce the same output if we used the exact waveform or we used the approximation of 10 to 90 percent. The circuit reacts little in the time between 0 and 10 percent, and little between 90 percent and 100 percent, in applications in which this convention is valid such as when measuring the delay time of a logic element.
For measuring slew rate, we approach the problem a little differently because it's a different kind of specification. We have more leeway, but we also need to pay attention to what is happening in the application itself and how the slew rate affects it. This is a more advanced topic though. To understand this better we could look at some actual waveforms and do a few little calculations to show what this is all about, but in the mean time you'll be ok to think of it in more simple terms.
One final note is that often in the data sheet they specify the percentages they have used to make the measurements. Often it is 10 to 90 percent but there is lots of room for variation here.