Just getting back to the pump vs flow thing. It gets intertwined with block design.
As stated a couple of times, weaker pumps benefit more from "system" resistance, meaning that 1/2" tubes and barbs are beneficial.
Now for Swiftech who can mostly control which pump is coupled with their blocks, especially for kits, this allows a little more stability in terms of designing a block to match the system. As we can see from various tests, pushing much more than 1.25GPM through an MCW600x yields very minimal gains, but this would not have been by accident. This is pretty much exactly the sorts of flow rates that the MCP600 in the 3/8" Swiftech kits would be pushing, so Swiftech can build a block to reach a point of rapidly diminishing returns given the pump that they couple with their kits.
Where it gets a little gnarly for "uncoupled" blocks is that people run with anything from an Eheim 1046, right up to the Iwaki's. So where does one "optimise" for now?
If the assumption could be made that everyone had a D4 at the least, the block could be solely focused on accelerating the water with that sort of pumping pressure in mind, and choke the flow rates to a point where 3/8" ID tubing setup doesn't play a hugely significant role in the system-wide flow-resistance.
If I were building a kit, I'd be selling a "D5", some 3/8" ID tubing everything, and the CPU block would be choking flow rates to the ~5LPM (~1.25gpm) mark, with all that internal block resistance being spent to accelerate the water to high velocities (~10m/s at the least).
However, if someone then comes alone with an Eheim 1048 (or a 1046), as certain "international" testbeds use, flow rates will plummet to performance sapping levels as the resistance of the tubing becomes fairly significant for the pump. So naturally we would need to do everything we can to assist the pump, such as fitting 1/2" ID tubing.
Where I'm getting at here is that 1/2" ID tubing is "required" only in the "chicken and the egg" sense - because many people used to, and many do still, use weak pumps - this necessitates 1/2" ID tubing to get the most benefit out of the pump (1C for $3 is a good deal) - which in turn drives low flow-resistance block design, which in turn drives the need for strong pumps to get the most out of those low-resistance designs - which in turn reaffirms the need for 1/2" ID tubing.
If everyone were using strong high-pressure pumps, we could all use 3/8" ID tubing, design the waterblocks accordingly, and the benefits from switching up to 1/2" ID would be more like 0.1-0.2C or so, or far less justifiable given the drawbacks to 1/2" ID tubing.
I'm not against 1/2" ID totally - it's just that it has evolved over time to become the accepted standard - but we need to understand why we are here using it, how it got that way, and not turn our backs totally on the possibility that it doesn't always have to be that way.
I then hooked them up in series, hold on to yer pants! I had a hard time totally restricting the flow with my thumb at the outlet tube! If it matters, my pumps say they were mfgr'd in 06/04, so just a month ago and they are only 60 places away in the serial #'s.
