42-AI
Module 0X - subject / scale: Discussion about the data type of the outputs
- Module: 03 and others
- Exercise: scope=all
Some evaluators are very strict about the data type of the result. But there is no mention of a expected data type for the outputs most of the time.
Thus, we can:
- [1] decide for all the outputs the expected data type
- [2] specify in the subject + scale that as no particular data type is expected, outputs should be either float or int (easy way).
As we wish to have a clear version, the fastest solution is to specify that no specific data type (either float or int) is expected. What do you think ?
IMO: I don't think asking for a data type is necessary. it is something that will be done in time naturally. Once you reach a bit more complex exercises, like MyLinearRegression, you need to make you code work, so the students will be constrained to use correct data type I think. AND if they really wish to use string each and every time, then their code will be slower, and it will be at some point an issue they will learn how to fix. No need to teach them, it is a lesson they will teach themselves.
