In "The New Organic Grower", Eliot Coleman describes a real-world approach to crop rotation. (There's a full chapter on it.) He works through an example of a theoretical 10-year rotation, and discusses the real-world 8-year rotation he's been using for a couple of decades. You could scale it back to 3 or 4 year if you wanted to.
First, divide up your crops into groups along two or three different axes: botanical family, cultural requirements, space required, etc. Note that you will have crops that land in different groups -- turnips+carrots will be together in a "root veg" group, but turnips+broccoli will be together in the "brassica" group.
Then decide how many rotations you're going to have and get that many index cards (or scraps of paper).
Create your rotation groups by putting together the vegetables you're going to grow together. Use some of the groupings you created above to guide you regarding which to put together since they share cultural requirements (planting in spring in the warm end; but see below too). Using the example from your question, you'd put salads, radishes and carrots on one card. Since you grow very few brassicas, you may put brassicas on a card with your herbs and onions. (I'm just making up a -- possibly poor -- example, assuming this will fill up the space allocated to a rotation.)
Shuffle your cards around until you have a good set of preceding crops. E.g. perhaps you want your peas before your brassicas because you've seen that peas help brassicas and onions. Or (using an example from the book) you put potatoes, then squash, then roots -- because the culture applied to potatoes and squash reduces weed pressure, and root vegetables don't handle weed pressure well.
Be willing to experiment. In my garden I've used a 5-year rotation, but I'm going to adopt a 7-year rotation starting next spring. I also want to shuffle around some of the sequences I have -- for example, I've learned that it is rather inconvenient to have cucumbers located next to tomatoes. Two space-hogs sharing space makes harvesting a hassle.
The uneven distribution of light is a hard thing to fix. Maybe you can't fix it. And maybe it doesn't matter -- experimenting a bit will help you find out.
Consider green manures and fallow periods. Coleman also discusses this in a chapter of the book. As part of my 7-year rotation, I'm planning on leaving about 1/4 to 1/3 of the space to a green manure for a year at a time. (I'm starting red clover late this summer that will grow in that spot all next summer and get turned under next fall.) This is an excellent path to soil building. (He also discusses -- and I'm considering -- pasturing chickens and/or sheep on the area to add fertility, but that's out of scope for most people on this site.)
As far as an uneven distribution of warmth and soil quality goes, part of the point of the rotation is to give you a chance to even these out. Use the rotation to improve the soil quality in year 1 for the crop that will grow there in year 2. E.g. plant peas to add nitrogen for the following crop of brassicas, and be sure to capture the nitrogen using a cover crop of winter rye that you sow immediately after removing the peas.
You can help the soil warm up in spring by doing certain things in the fall. Forage radish, for example, will winter-kill, leaving holes in the soil. These are supposed to dry out and warm up earlier in the spring. If you normally spread a heavy mulch, or plant a thick winter cover crop, don't do that before the "plant in early spring" part of the rotation -- it delays soil warming. If you normally plant "on the flat", then in the fall mound up some beds for planting into in the spring -- these will warm up faster.