Just so we’re clear on this, i’m not saying that every private method should become a public method in a new class or anything like that. The backward compatibility burden you’re talking about is more of an issue when you’re writing reusable libraries or frameworks. In that case you could just make the ‘new’ classes internal. But i don’t think it’s that big of an issue when writing applications.

but like i said, if you feel the need to specifically test something that is implemented in a private method of a class, it’s usually a strong indication that something is wrong in the design. In a lot of cases, it really does mean that there is a new class just waiting to be discovered. And in those cases, who cares if people start using the functionality that was once in a private method but is now in a new class? If you did it right, the functionality of the new class is now properly reusable… so is there really a problem with people reusing that functionality?

After a while when you decide to change OrderLineParser class you can find out that InvoiceParser, ReceiptParser, BOLParser and other document parsers unexpectedly changed their behavior too.