At the beginning of 2023, the IBEX probe sent into space to analyze the outermost region of the Solar System stopped communicating with Earth and NASA technicians had only one solution left: turn it off and on again. The command to reset worked and allowed IBEX to resume its activity, confirming once again the usefulness of turning devices off and on again when they stop working as they should not only in orbit, but also here on Earth.
Turning it off and on again is almost always the first advice that computer support services give to their customers, and it is also usually the first thing that children say to their parents who are having difficulty doing something with their smartphones or computers. For those who have a minimum of familiarity with electronic devices, it is the most banal and often effective suggestion that can be given, while for those who have little familiarity with computers it almost sounds like an esoteric practice: perhaps because in the eyes of less experts an object so sophisticated and with a mysterious functioning it cannot be fixed with such a simple move.
Yet, turning it off and on again works in many cases and not just for the devices we have in our pockets or on our desk. Trains, ovens, cars, hospital monitors, airplanes and – as we have seen – space probes have computers inside them that are more or less advanced depending on the case, with programs that make them work thanks to the instructions contained in their code. Whether they are simple and rudimentary or extremely complicated, they can all contain errors or programming gaps that in some circumstances cause malfunctions. Sometimes the problems resolve themselves, for example by closing the program that caused them, but in other cases they continue to slow down the computer by making it carry out useless work and occupying large portions of the temporary memory (RAM), on which part of the of active processes.
The causes of a malfunction of a program that extends to the entire operating system (such as Windows or macOS) of the device can be many and different from each other, but it is important to remember that in its essence a program is a more or less elaborate set of rules that start from the concept of if this happens then do that (with artificial intelligence systems things are a little different). In the event that there is some error in the sequence of tasks that a program must carry out, for example there may be a “memory leak”, with an anomalous consumption of memory which is not freed at the end of the program. execution of a process.
After a certain period of time, available memory runs out, resulting in program or system failure. Once upon a time the problem manifested itself immediately because when the available RAM ran out there wasn’t much possibility of intervening, the device didn’t respond at all, while today many operating systems use a system to continuously transfer part of the RAM into fixed memory, in way to free it and use it for other processes. The most rudimentary devices are not equipped with these systems and rely more on temporary memory, one of the reasons why they only work again when you turn them off/on, restoring the initial conditions, when the memory was not saturated.
To get an idea of how a memory leak can occur, the example of a program designed to operate an elevator is often given, the code of which involves these actions:
• if a key is pressed for a certain floor -> recover some memory to keep note of the floor number and insert the floor number into memory;
• if you are already on the requested plan -> there are no actions to do, so end the program;
• if you are not already on the requested floor -> go to the floor indicated in memory;
• if you have reached the requested floor -> empty the memory used to remember the floor.
In this example, the memory leak occurs if the button for the floor where the elevator is already located is pressed, because the instructions do not provide for the memory to be emptied: the code says to do nothing, because you are already at the plan, but the command to remove plan storage is not required. A memory leak of this type would not have an immediate effect, because the elevator would probably have enough memory to handle multiple requests for the floor it is already on before running out of space, but sooner or later that condition would occur and would prevent the elevator to function normally. In that case, to empty the memory and restore the functionality it would therefore be necessary to turn the system off and on again, in order to eliminate the information in memory and make it available again.
The elevator example is of course an extreme simplification of the much more complex processes and conditions that lead to a memory leak, but it helps to get an idea of what the benefits of a system reboot could be. Naturally in this case, turning it off and on again restores the normal functions of the lift, but does not resolve the problem due to the programming error which may therefore be repeated again after some time. To solve it, you will have to intervene on the code, adding the correct instruction to empty the memory when the elevator checks that it is already at the requested floor.
Something similar, here too with greater complexity, also happens with smartphones and computers that now make the world work. This is why their operating systems and programs are periodically updated: the most recognizable updates add new features, the less obvious ones resolve programming or design problems that are gradually discovered either by those who develop them or by those who use them (or often from the programs themselves, which automatically send data when they crash, from which information can then be drawn to fix them).
The causes of a malfunction can be many, they vary from device to device and it is not always sufficient to turn it off and on again. Compromised files or physical component problems may prevent functionality from being restored and require other interventions. In short, turning it off and on again is not the solution to all IT problems, but it is almost always the first thing to try.
