I have seen there is a feature called pressure advance which aims compensating filament flexibility especially for bowden setups.
There is a feature in Marlin which is in some way similar to this but compensates variations in line thickness caused by acceleration: linear advance http://marlinfw.org/docs/features/lin_advance.html.
I think this is different and if so, I would love to see it in the Duet3D RepRap because it works really well in Marlin.
The pressure advance algorithm that RepRapFirmware first provided 3 years ago has always been linear. Marlin "invented" the same feature 2 years later and called it "linear advance". Once difference is that in Marlin the K factor is expressed in motor steps x seconds, so it is sensitive to the extruder steps/mm setting and will change if you change extruder microstepping; whereas RepRapFirmware expresses it in seconds so that it is independent of extruder steps/mm.
Ok, so this is essentially the same but even better? Great!
What is best practice to configure it? Is there a form to create a test gcode like it exists for Marlin?
There are a couple of threads in this forum that linked to test files for calibrating pressure advance. One way of to print a cuboid in spiral vase mode and adjust pressure advance on the fly using M572 to get the best corners. Bear in mind that if you use high values of pressure advance, you may need to reduce retraction, especially if you are using an all-metal hot end.
One question about the D parameter for M572:
The wiki says D is for the "extruder drive number". Is this the drive number or the extruder number?
In other words: What should I use for the first extruder (E0)? D0 or D3?
Edit: Just tested it. D3 throws an error (invalid extruder number), so D is the extruder number (D0 for first and D1 for second extruder). This should be made clear in the documentation.
The D numbers are the same ones you use in the M563 tool definition command.