Add support for uncertainties of measurements

[davidclemens]

Feature description

Add support for quantity uncertainties. This can be done in several ways:

1. Create a new double subclass that supports uncertainties. See subclassing built-in classes and defining interfaces in the MATLAB documentation.
   • Pros:
     • error propagation is an inherent feature of the class
   • Cons:
     • increased memory footprint
2. Track uncertainties in DataKit.dataPool and make use of it in methods that need to propagate an error.
   • Pros:
     • low memory footprint
     • normal computation speed
   • Cons:
     • error propagation has to be implemented every time the values are used

• Implementation examples:
  • uncertainties + units + error propagation
  • units
  • error propagation functions
• Inspiration
  • pint (fully featured Python unit implementation)
• Resources
  • Dimensional analysis on Wikipedia

    Problem and motivation

    Uncertainties are currently not properly supported both in storage of data (e.g. DataKit.dataPool) and also during calculations as error propagation but are an essential part to real-world measurements.

Is this something you're interested in working on

Yes.