One thing that might be very informative is: “look at the IPO window.”
(There is actually a “screen” for that, which you can switch to. It’s simply a pre-fabricated arrangement of windows and settings…)
Anyhow: “IPOs are ‘how things work’ in Blender.” When you, say, set a series of keyframes, what Blender does is to place specific markers on a (probably Bezier-curve based) IPO curve.
The actual position (say…) of the object at any particular frame is determined by three separate curves (LocX, LocY, and LocZ). Each time you make a keyframe, you place another set of markers on those curves, at a particular frame-number. The curves always pass directly through each marker; the “in-between” frames are then determined by interpolation.
As you will see when you fiddle-around with the IPO window, there are many groups of IPOs available. Yes, everything in Blender, with respect to “time,” is determined by IPOs.
When you see a character’s movement side-by-side with the actual IPOs that drive it (as you can do in the screen aforementioned…) a great many questions should be answered. This is how the “ease-in/ease-out” of standard motion is arranged; this is also how you can change it. Editing an IPO-curve is done exactly like editing any other kind of curve.
My experience was that Blender’s behavior seemed erratic and nonsensical until I saw the actual IPOs and glommed-on to how they work. Then, it suddenly seemed “elegant.” And when I realized just how many IPOs are out there and in just how many ways they can be used, it became “whoa-a-a-a-a… cool!”
(And, “from one software designer to another,” I still feel that way… My hat is still off to whoever-you-are-out-there.) (Ton?)