I'm an electronics engineer too (but working in IT).
From a control systems point of view, putting the sensor away from the thermostat will inherently make the system less stable. The control of feedback i.e temp/open close valve is inherently more stable when they are together with little/no transfer function (i.e. that created by the large reservoir like a radiator). thinking intuitively, having the thermostat and fan switch working in sync is a more stable solution as it all works off water at the same temperature at the same time.
Controlling a fan speed is simple in terms of a PWM controller. Problem is some fans are 250W (20+ amps) upwards requiring some beefy (technical term) transistors. It also gets awkward fabricating circuit boards for these. And the potentially harsh conditions in the engine bay etc. are are problems. You can buy kits to do this should you want to, it is probably pointless (except as a hobby exercise) to engineering your own solution. Certainly microcontrollers are far, far too sophisticated for this application. I'd be looking at NE555 and adjustable pot technology at most...
In the real world most car manufacturers don't generally do this anyway (or at least most that I looked at). If you look at the specs of many thermoswitches, they have three terminals and trigger at two different temperatures (for example those used on a VW). These are great as you can either select a temperature to turn the fan on or engineer a solution for a 2 speed fan. Certainly on modern fords there is a simple (but substantial) resistor pack mounted behind the radiator as a speed control which I guess is switched from the engine block temperature sensor (rather than a switch directly) via the engine management.
So basically for refinement what you want is a slow speed which should be enough to cool in most conditions (most likely stationary/traffic and no fan when moving) and a full speed that only kicks-in in extreme conditions such as summer traffic jams with the aircon running.
In terms of "variable cooling" your stat does this to a degree as the temperature quoted is when it starts to open, it isn't a simple on/off. Full flow only happens another 7-8 degree higher when fully open.
Another thing to consider is that the thermoswitch would need to be set to match the thermostat. You'd want the stat to open first and the low speed cooling kick in say 5-7 degrees higher and the full speed say up around 100-105 degrees. In old cars it's not uncommon to use winter/summer stats; this far from optimal with electric cooling, unless you can change the swtiching points to suit - most likely you'll otherwise have the fan spinning pointlessly with the stat closed. With a good and stable cooling system you'll probably want to run the higher temperature winter stat for efficiency and performance all the time (who the hell changes stat on a modern car?). In other words instead of the 82 degree most people use you might want to look at running the 88 degree. Stability at higher temperatures is also why modern cars tend to have more substantial cooling systems than the marginal ones on our cars even though modern engines are inherently more efficient and waste less heat anyway - certainly in traffic jams my old Audi diesel couldn't make enough waste heat to heat the cabin in winter even with the cooling system completely shut off.
In conclusion I think the engineering consensus is you don't need a sophisticated solution, except perhaps in very new cars where temperature is far more important in terms of emissions control etc. I'd run a warmer stat and rig a two speed solution with a large single puller fan controlled from the top hose.
You'll get better fuel economy, not as most of these discussions usually mention just by NOT running the cooling fan from the engine all the time and the resistance this adds, but you'll get quicker warm-up and likely get more efficiency by the engine running at the right temperature more of the time, especially in winter. And of course less wear and generally cleaner running. A positive sign is you'll generally need to lower your idle speed right away. The low 80 degrees are really too cold even for a 1950/60s design IMHO; this is more a compromise to err on the safe side because inherently an engine driven fan is an awful solution; too little cooling when you really need it in traffic and too much when you don't when on the move.
e.g. This one triggers in the ranges 95-84C/102-91C which I'd guess would be a good match for a P6 running a 88 degree stat.
http://www.vwheritage.com/vw_spares_The ... tartPage=1
Rover themselves (perhaps ill-advisedly) actually took this to a further extreme by running the V8 (much) hotter still to keep the old engine within emissions limits into the 21st century.
Well that's a bit more than 2c worth. This is what I plan to do with my V8 anyway.
and not forgetting the inclusion of an SMCJ30A eh Vin 