Unless you already have some very specific jobs in mind I think it more illuminative to turn the "what to you want to do with this machine" question upside down and ask "what can't I do with this machine". "Can do" questions tend to rapidly concentrate on a few specifics. "Can't do" helps with the wider picture. One thing is certain. Once you have the machine you will be doing things with it you never expected to do or planned for.
If you aren't awfully careful, experienced in such choices and, frankly, a bit lucky the "can do" question may well prove to be a good way to screw up by the numbers! Despite impeccable logic. Did that with my first mill which proved to be a spectacularly wrong purchase. Second one was better but still less than ideal. Then I got a Bridgeport, which I can live with, 16 years and counting…. But I do home workshop jobs in 12 inch to the foot scale not model engineering so size matters much more to me.
Given that a Bridgeport can handle any sane job you can lift by yourself, and a few insane ones too, that's probably your best benchmark.
Being realistic the ram movement and head nod / tilt facilities are very much in the nice to have around just in case class. I've never nodded my head, have tilted it twice and moved the ram three times. But I have both tilting and titling / rotating vices along with a dual axis sine table. The Bridgeport has always had enough vertical daylight to accommodate one of those so I've little need to move the head.
Probably the major pitfall in assessing machine capability is underestimating the amount of space needed for work holding.
My Bridgeport has plenty of vertical room to take tilting devices for angled work. Smaller bench top machines don't, so if you expect to be doing angled work you need to consider what can be done with the head tilted. My second machine, a Chester supplied square column machine similar to the Warco Super Major linked to by Thor, had what appeared to be a useful tilt capability. But the table was significantly shorter than the Warco one with less travel so the job often couldn't be moved far enough to the side to accommodate the tool offset generated by the tilt. Especially with larger drills.
Rotary tables are another pitfall. My Bridgeport accommodates 10" and 12" tables easily enough giving plenty of room to clamp jobs on using the usual Tee slot devices. Smaller machines are generally limited to 8", 6" or less tables with a lot less surface area so clamping takes up disproportionately more space. An 8" table has only 2/3 rds the area of a 10" but twice the area of a 6" one. For 6" and less tables I consider a grid of tapped holes plate better than Tee slots as giving flexibility for more, albeit smaller clamps. (Mine has M6 holes in 5 spot dice pattern relative to 25 mm squares. Still comes out occasionally for teeny jobs.)
Its also important to ensure your chosen machine doesn't have features that, however illogically, just happen to drive you nuts in a "fingernails on the blackboard whilst jumping on your favourite corns" way. I seriously disliked the big rectangular head on my Chester supplied machine because it not only greatly obstructed visibility when working with short cutters but also seriously cramped setting up jobs. A longer table would have been a major help there giving space to move most jobs clear of the head.
Got to the point where I seriously considered cutting the corners off the head and turing the column sideways with a ram style mount just to get it out of the way. The back of my envelope said the useful X travel would be little reduced, albeit offset. Never underestimate the ability of frustration to drive you into over-reaction! Then I got sensible, and a redundancy payment, so a (much) bigger workshop got built.
Clive