Imagine a future where people would look back and be puzzled why we could not see the correct application for computers and computation. Perhaps in a similar way we might be puzzled why technology arount the movement and control of water was not applied to domestic sewerage immediately.
As a computer enthusiast I am looking for new and interesting ways to apply comuter technology to the real world, particularly in the form of some product that is worth enough to people to keep me comfortably employed.
So you pick a random industry and ask: what is the correct application of computation to that industry. Lets take an example: commercial aviation. Examples of the correct application of computation:
I would say these are the main areas in commercial aviation where computation provides the most benefit. Computation is not at the stage yet where the job of a flight attendant could be handed to a computer (a robot) because the computational load of navigating a human environment is handled more efficently by people than a computer.
Control and Simulation are self evident scenarios where computers excel. The other area is in paperwork (inventory and scheduling). What are these things? Arguably inventory keeping is the origin of writing, and there is a proven track record of inventory keeping being valuable. Most complex civilizations have inventory keeping as a core feature of their inner workings. One of the earliest applications of commercial computers was the automation of various record keeping dutues.
Why is it so important to have record keeping? If I loan my neighbour five cows why is it so critical that I put five notches on a stick representing the cows when the five cows are right there to stand for themselves? Why fill jars with olives and then record somewhere else how many olives you have in the jar, when you can just open the jar and see? Could this be why it took so long for record keeping and writing to happen - perhaps it just wasnt needed!
Could it be the same for computers? Perhaps we just dont really need our lightbulbs to be controlled via an app, when a switch on the wall is so much more secure, convenient, bug-free, intuative, repairable... the list goes on.
Control and Simulation are areas where computers excel, there is little doubt of the usefulness of applying computers has massive advantages. Self driving cars are shit now (and probably will be for a long time) but more mature vehicle control assistance technologies such as ABS braking and fuel injection have proven their worth many times over. Computer controlled assembly lines have completely revolutionized manuafacturing. These applications tend to be embedded.
Simulation is a key area where a lot can be done above and beyond what people can do manually. Simulation requires modelling the real world in some current state, projecting that into the future in order to be able to predict future events and then exposing that to people to that they can prepare. The human mind runs simulations all the time to be able to predict what the most sensible course of action is. Simulation can mean climate simulations so people can prepare govenments for global warming policy, flight simulations which can prepare pilots for real flying and economic simulations that can show projected growth of a company and effect hiring decisions. Simulation is a remarkably varied and valuable area of computing.
How do modern trends fit into this model? Collect massive amounts of data and analyze for models. Machine learning models are black box simulations used for their predictive power. Big Data is also about enabling simulation in the broad sense. Control software is still mainly embedded and PID loop based. Advertising and user tracking is also about simulation to predict display ads.