I have been creating mock-up fixtures for LS2 results and question types
these days. After that , i encountered some problems regarding the design of
the question structure , particularly in Model Answers model.
As far as i know, a question is composed of question text and a set of
possible answers ( predefined or not predefined ). In the current design ,
model_answer acts as a central hub to store the answer. That meant we mixed
the labels ( as in LS1 called ) and answers together
For example :
Apple ( this is a label ) ------- __________ ( this is an an answer )
This is fine with the simple question structure , however, when the question
structure get really complex with the some question types having multiple
labels' type ( like multi flexi , we are actually mixing the labels and
answer together, treating them all the same. The only thing that separates
labels and answers is category. Since we can have as many categories as we
want, it's possible to assign all predefined answers to a category and use
Apple: -- category: label
Answer: -- category: answer
The second way to approach it is to use different models for possible
answers and labels ( like LS1 which have labels and possible values ).
Please give your thoughts on this.