What happens with your tectonic plates and the resulting mountain building action is going to depend on a couple factors, most importantly the direction of drift (which in your example, you've only listed A as moving NW, although you do mention that A and D are pulling apart, so I'm assuming D is moving SE.) The other big factor with plate boundaries is whether you are dealing with continental or oceanic crust. Generally plate boundaries fall into one of the following categories:
Rifting Zones, where two plates are moving apart. Magma comes to the surface as "new" land is formed creating mountainous areas. Best examples of this are the mid-Atlantic ridge (longest mountain range on Earth) and the Rift Valley system of eastern Africa.
Sliding Plates, where two plates are moving past each other. You'll get a lot of earthquakes here but not necessarily a lot of mountain building. However, the stress caused by the strike-slip action can have any number of effects on the plates involved, and you can feasibly get isolated areas of volcanism and mountain formation.
Subduction Zones, where two plates ram into each other head on. Again, generally continental crust is going to get thrust upward as the oceanic crust subducts beneath it. As the subducting plate dives into the mantle, the high water content of the oceanic plate causes magma to form at the mantle-crust border. This superheated magma rises, leading to volcanism and mountain building at the surface. This type of mountain building is exemplified in the Pacific Ring of Fire, as well as the Alps and Himalayas. Another type of mountain building associated with this action takes place on continental plates as it "crumples" and uplifts when it goes head to head with a neighboring plate. The Rockies of North America are an example of this action.
Of course, not all mountain ranges need necessarily be at active plate boundaries, even if you are striving for realism. As Hai-Etlik mentioned, you can have hot spots creating volcanic island chains in oceanic plates or supervolcano ranges in continental plates, such as Yellowstone. Some mountain ranges, such as the ancient Appalachians, were created by subduction of tectonic plates that have long disappeared into the mantle. (The modern Appalachians were created through uplift, much like the Rockies.)
So basically I'd keep playing around with your plates, labeling them with different directions, until you get the right combination of mountain-building action. When creating my world I started with the plate boundaries, continental shields and direction of drift and then sketched out the resulting land masses. But there is no reason why you can't do the same thing in reverse.
Your moving glacier does present a bit of a problem. While my expertise on glaciers isn't as thorough as my knowledge of plate tectonics, I will try to offer some observations. While glaciers have the ability to change morphology of mountain valleys, the affect is going to be pretty localized (the longest glacier in the Alps is only about 20 km long). If what you are really talking about here is the expansion of an ice sheet, then that, I suppose, could possibly reach as far south as the 30th parallel, but bear in mind that that's going to be accompanied by major climactic repercussions (and I thought
MY world was cold). IIRC, the D&D Forgotten Realms world of Abeir-Toril had a huge "glacier" running amok through the center of it, but I believe that was a magically driven phenomena. *Correction: it was a desert running amok, not a glacier, although parts of it did have snow on it. My bad.