Voltage references are a humble piece of hardware, their sole function is to provide a stable, known voltage. This constant, known value of voltage can then be used as a reference for ADCs and DACs as well as provide a precision current source.
I recently got hold of an Analog Devices AD587KN high precision 10.000V reference chip.
This model of chip has an output value of 10.000V ± 5mV (that is, an output value of 9.995V to 10.005V) straight out of the factory. A voltage drift of 10ppm/°C at 25°C meaning that the output voltage will drift by 10μV for each 1°C the chip is exposed to. Additionally, the chip has a voltage trim input, so if you have access to a precision voltmeter, the chip’s output value can be adjusted even closer to 10.000V.
Alternatively, the chip’s output can be trimmed to a value of 10.24V. You may think that a value of 10.24V seems like a strangely familiar number. A value of 1024 is the decimal representation of 10bits, that is 2∧10 = 1024. Why would I want a voltage reference that outputs a value of 10.24V? Because it makes any ADC or DAC conversions much simpler.
For example, consider that you have a 10bit ADC and you provide it with a precise 10.000V reference voltage. Your 10bit ADC has a resolution of 10.000V/10bit = 10.000/1024 = 0.009766V/bit. Or more simply, the ADC can resolve 9.766mV (3decimal places) per bit. In any software that is manipulating data from the ADC, a scaling factor of 9.766 has to be used and it is apparent that this is a cumbersome number to deal with (let alone your software will be slowed down by having to perform floating point calculations due to the decimal point).
For more detail: Analog Devices AD587 10V Reference