Perhaps not the smartest specimen for Lynch, but I wanted to start somewhere)
For the beauty of the code and the convenience of writing modules, implemented classes for working with individual information entities. For example, such a thing as an interface language has its own ID, and for most operations this information is sufficient. At the same time, there are rare situations when you need to get additional information about it (RFC-code, for example, which can be found by referring to a separate table or file). Without much hesitation, I set the initialization of variables to the class constructor ... The problems started when I moved to a machine with a slowing SQL server (a very useful exercise, as it turned out) - the test machine fell into a timeout with completely harmless operations. Not comme il faut — I would like to use processor time and SQL connections more efficiently :) But I also didn’t really like to use special methods for loading such as Fill (). As a result, I came to the simplest (in my opinion) solution of the available :
public class { private String ;
public String { get { if ( == null ) { = (); } return ; } } }
')
Thus, practically no changes are made to the code (when creating a property with the name of an old variable, and then renaming the variable), and in some cases, the excess load is reduced to zero. But doubts about rationality still remain.