I've started to ignore the recommended locations more often. I'm not sure what algorithm the AI uses to select city locations, but they put a lot of them in what I consider bad locations. That said, I don't have a "rule" so much as several rules, depending on what type of city I'm interested in founding and what kind of resources are in the area. A couple of basic principles that I follow (currently working on Prince level):
1) I don't worry about using every tile, even every resource tile. I'd rather make each city I do build in a solid location and pick up resources outside the BFC through border pops than pack the cities in. The reason for that comes down in part to maintenance costs. The two factors that influence city maintenance are distance from the capital and number of cities. I accept slightly higher costs related to distance since that allows each city to be more productive than if I were to put out more cities (so more maintenance from that source) and cover every tile, even the poor ones.
2) Every city location needs both food and hammers to grow. Food resources, flood plains, and grassland all work for the first part. Hills and mineral specials for the second part. The mix of the two determines what type of city I'm building. Lots of food makes a site good for cottaging (so, lots of research and gold potential) or specialists (lots of Great Person Points, plus customized city output depending on which specialists [but usually scientists]).
3) To get my hands on key early resources like copper, horses, iron, marble, and stone, I'll put cities in some pretty crappy locations. Those resources are worth the maintenance burden of a small city, IMO. Otherwise, unless I'm building a city to block AI expansion, I normally won't build it unless it has either a couple of food resources/flood plans and/or several grassland tiles so that whether I'm building a production, specialist, or cottage city it will get big enough to do me some good in the long run.
4) Settling in the middle of a forest area can be good for chopping out wonders or units, depending on what you need. Don't be afraid to cut down those forests.
As for maintenance, as the difficulty level goes up it becomes harder to pay for the cost of numerous cities. Some observations:
1) The 60% research rule is a decent one in the early game, unless you are involved in a war. Later in the game, you can maintain decent research rates through buildings, scientist specialists, and settled great scientists. Couple those with espionage and tech trading and lower research rates are fine.
2) I don't usually build more than 3 or so cities before I'm trying to take one or more from my neighbors. If I like the location of the city I take, I'll keep it. More often, though, unless it is a capital or a shrine city I'll raze it. Later in the game, I'm more inclined to keep the cities I conquer rather than raze and settle new ones. You can fund a lot of deficit spending while expanding during a war by conquering cities.
3) Don't neglect financial infrastructure and trading spare resources to AI civs for cash.