Skip to main content

User login

What is not?

 Impossible is not a fact, its an opinion 

Required evidence

May 23, 2010 by dangyogi

Required evidence works two different ways thinking about it:

  1. This question must be answered favorably.
  2. This question must not be answered unfavorably.

But from algorithmic point of view, they are the same.  In either case, there are three possibilities:

  1. The question has been answered and the requirement is met.
  2. The question has been answered and the requirement has not been met.
  3. The question has not yet been answered.

The proposed method for this works as follows:

The question has been answered and the requirement is met

This is the most puzzling situation.  My first suggestion was to simply act as if this question was no longer on the list of questions.  The reason is that this allows the overall CF to reach 1 when all of the other questions are answered favorably.  The drawback to this is that when none of the other questions are answered, or they are all answered unfavorably, the overall CF would be 0 -- thus discounting the evidence of this required question.  Does this sound OK?  Any ideas?

The question has been answered and the requirement has not been met

Since this question is a requirement, the overall CF is simply forced to 0 in this case.  The answers to the other questions don't matter.

The question has not yet been answered

For this case, we go ahead and assign a CF to the required question.  When it is not answered, this CF gets added to the total divisor for all of the other questions, thus preventing the overall CF from reaching 1 in this case.  But the other questions can still contribute to the overall CF, thus suggesting this diagnosis as a possibility.

 

Premium Drupal Themes by Adaptivethemes