Recently when working on Observability at Blinkit I realised the importance of continuity and smoothness. I am trying to actively apply this to multiple different areas now.
in an ideal world I think that most of the critical parameters that are plotted against time or any other similar parameter that is plotted in X axis should be continuous and smooth. Wherever it is not continuous I think we can look at it like a downtime of a system and whenever something is not smooth we can look at it as a failure in forecasting the demand or and in efficiency in the market.
For example if you take the earning potential of a person in the four years of their graduation it is generally at a zero value but the moment they complete their graduation and get a job suddenly their earning increase so it is more like a step function. Since this is a non smooth function it becomes difficult for the students to get feedback and get to know whether they're eligible for the jobs or not.
One way of improving this is by converting the step function into a gradually increasing function (smooth flow function) or at least break down into multiple steps so that the height of the step function is reduced and it is converted into a staircase. Looking at it from this perspective internships are great way to reduce this risk and also reduce the height of the step function. Though the amount that is paid to students during the internships is way below what it is paid to a full time engineer, it is still a great way to give students an insight into how industry ready they are and also to impart critical skills for them in the actual jobs that they will end up working in.
The way to reduce this will be to convert this single step into a series of staircases which are smaller steps to climb. The best case scenario will be a smooth curve and this happens when the height of each of these individual staircases is almost close to 0. One way of doing this would be to come out with a series of smaller activities that students can perform to earn smaller incentives.
I am trying this approach with mentoring 150 people and open micro tasks repository experiments