Because they offer the most capability for the money, and everything around them is cheap and available. The operator can see and reach the work, setup is quick, tooling and workholding are mass-produced and inexpensive, programming is familiar, and almost any machinist can run one. For general work, short runs and prototyping, nothing else competes on total cost of getting a part made.

Chips fall into the work rather than away from it, so they must be actively cleared or they recut and shorten tool life. And a three-axis machine can only reach one face at a time, so multi-face parts need multiple setups, each consuming time and introducing the possibility of positional error between features machined in different setups. Both limitations follow from the vertical geometry rather than from build quality.

A rotary table that indexes the part between faces, or rotates it continuously for cutting around an axis. It converts a large share of multi-setup jobs into single-setup jobs, which removes the accumulated error between setups as well as the labour. For most shops it is a far cheaper route to multi-face machining than a five-axis machine, and it is the usual first upgrade once three-axis setups become the bottleneck.

From the largest part you realistically expect, plus the fixture that holds it, plus clearance for the tool and holder. Buying to the exact part size leaves nothing for workholding, which is the usual regret. Check table load capacity as well as area, since a heavy fixture plus part can exceed it, and confirm that the Z travel accommodates your longest tool with the part on its fixture.

Not by itself. What matters is torque and power at the speeds the work needs. Aluminium at high spindle speed with small cutters is a different requirement from steel at low speed with a large face mill, and a machine optimised for one is compromised at the other. Read the power and torque curves. Rigidity of the structure often limits achievable cutting rates before the motor does.

A C-frame carries the spindle on a column at the rear, which is compact and economical but allows some deflection under heavy cuts at extended reach. A bridge or double-column design supports the spindle between two columns, giving greater rigidity and better accuracy over larger travels, at higher cost and with a larger footprint. Choose by part size and the tolerances required at the extremes of travel.

By solving chip handling first, since chips are what stops unattended running on a vertical machine: through-spindle coolant, wash-down, a chip conveyor and an enclosure that actually drains. Then tool life management with sister tooling and load monitoring, in-process probing to catch drift, and adequate coolant and lubrication reserves. Unattended running is usually limited by chips and coolant rather than by the control.