I try to follow the principles of SOLID . But every time it comes to user interfaces, I find that there is friction between the clumsy screens of hybrid, aggregated data that the client requires and the good principles of shared responsibility.
Now you can divide and conquer the various pieces of a typical user interface into classes with one responsibility, but then you encounter all the interesting structural problems, because the so-called "separate" parts of the geom are often different representations of the same general condition or, at least partial representations of an overlapping state.
I often end up arguing with the rather clumsy controller classes for my views, which are not very similar to SOLID, but this is a rather controversial coding practice, and it bothers me a bit. It seems that the complexity of its separation is not worth it.
So how do you deal with this?
source
share