1. Flat top.
2. Flat top.
3. Flat top.
4. Flat top.
5. Flat top.
6. Flat top.
7. Flat top.
8. Flat top.
9. Flat top.
10. Flat top.
These are a few other things, but they are irrelevant if the top is not flat:
11. Convenient access for bit changes. (Convenient for you, the user. There are plenty of opinions on this.)
12. Convenient height adjustment. Convenient for you... I actually prefer an adjustment easily reached while crouching down to sight the bit from near the table top. Many don't.
13. On/Off switch on front of table.
14. Straight fence, square to the table surface. Tall and short fences are useful.
15. Some method of predictably fine adjusting fence position. Fancy fences are nice, but this can also be done if the fence is a straight stick held to the table with two clamps. All you need is a block, a 3rd clamp, and a feeler gauge. Clamp the block to the table with the feeler gauge between block and fence, then move the fence to the block.
16. DC. I like a shopvac pulling from the fence, since it has a small opening around the bit and good suction with small air volumes helps. A real DC works better in the table itself. I have a horizontal baffle between two supporting ribs, positioned about haflway up the router. The router protrudes through a hole in the baffle. The DC pulls air from above the baffle, below the table top. This collects well, and gives the router plenty of clean air, coming through in the correct direction (up).
You might notice that I have not addressed the main design questions, except to say what you need to be able to do. These things can be done, in different ways, with a variety of equipment. People have vehement opionions.
Should you use a router lift? They are accurate, convenient, and expensive. If you don't get one, you need to look at how easily and accurately your router can raise and lower. Many routers these days allow above the table height adjustment, though as I alluded, I don't care. Few have the clearance to allow above the table bit changes, which I think would be truly useful.
If you can pop the router plate out of the table, you can get at the bits easily enough. If you bolt the router directly to the tabletop, it's easier to keep the top from sagging because to the structural continuity, but might complicate bit changes. Tilt-top tables address this, as does use of a router that it quickly removed and replaced from its base.
One of the limitations of the Pat Warner bolt-to-the-table approach is that the table is usually at least 5/8" thick, so you give up some reach compared to a plate. That is probably why Warner always notes the extension distance when he reviews routers.
Pat Warner's books are worth reading, as are Hylton's. It can be quite interesting to contrast their advice, and look at how each carries out the same task.