Current state-of-the-art differential controllers for solar hot water are rather binary affairs. Wouldn't it be possible to combine delta-T sensing with circulator pump motor speed control to adjust water flow through the collector to maintain setpoint delta-T?
In my opinion, this would optimize gain from the system without using restriction valves to set maximum flow rates (you lose on bright days) or off when any "on" flow chills collector array below setpoint (you lose on hazy days). In effect, this would be maximum power point transfer for SHW.
Not disputing the idea - I've thought about it often, and long... but short of actually DOING it, am unsure the outcome is actually beneficial.
The greater the deltaT between the collector and the "cold" input water, the more efficiently you can make use of the solar energy. However a large quantity of "warm" water may not be sufficient if you need "hot" water. Given you can mix a smaller amount of "hot" water with "cold" water to make warm, the maximum efficiency may be to keep the water in the collector as cold as possible. Or, to bring in the coldest water you can until it's "hot enough" to be useful, then dump that into the tank and get a fresh batch of "cold".
I originally intended running my pump with variable flow as you describe, but the simplicity of a "variable duty-cycle batch heater" over-ruled, and has been running for several years now, very well...