The diagram you have posted shows the two possible alternatives.
The upper part of the picture shows a device connected to IN-0 via the two-core shielded, twisted pair cable, and that device is capable of driving the 4-20mA signal current that the IF16 needs. By inference, then, this device (which isn't shown on the diagram), is powered from another source, most likely it is powered from 120Vac or 24Vdc. This is known as an Active device.
The lower part of the picture shows a device that receives its power from the 4-20mA loop that the PLC input is monitoring. Such a device is classed as a Passive device, or "loop-powered". Obviously a loop-powered device itself must not consume more power that would exceed the 4mA minimum signal current. So assume the device itself requires 2mA to power the device's electronics, then it will regulate additional current through itself to "signal" 4mA to 20mA for the working range of the device.
In either case, the transmitter device needs power. Active devices just drive the 4-20mA signal through the analog input module with a current generator, while Passive devices "steal" their power supply from the loop, adding additional current to provide the required 4-20mA signal.
In the diagram, the "User-Provided Loop Power" must be a "stable" power supply, (usually 24Vdc), which must be provided for Passive 4-20mA devices.
It is generally not recommended to use the same power supply that powers 24Vdc Inputs and Outputs, although many people do this. It all depends on the capability of the power supply to maintain a constant voltage with a variable load, and the quality of the loop-powered devices to reject or filter out line supply noise.
So typically a control system will have separate 24V power supply modules, one for powering digital inputs and outputs, which needn't be that good at line/load regulation. And a second for powering "loop-powered" analog transmitters, which would be a much smaller supply with good line/load regulation specs. This, IMHO, is "best-practice".
