Interesting point, Konstantin. I think it is a matter of perspective. Part of the reason that I use composable services/components is their ability to process streams of data efficiently.

Every composable service is a reactive, actor model, message processor and can be fed a data stream for processing. Because they are re-entrant and stateless there is no need to instantiate a new service object for each message.

One serious problem I've been trying to solve over a 50-year career is a human one. What software models can best enable teams of human beings to translate the application knowledge of domain experts into working software that can be maintained, deployed, managed, and modified efficiently and effectively. Software that is self-organizing, self-monitoring, and self managing.

If you have the time, I'd recommend browsing the reading list at the end of the OOP article. It may not be the only way, but it does work.

-Dick-