Open Closed principle in design patterns

I have following options

• Modify the PrintTicket() method in Ticket abstract class. But this will violate Open-Closed principle.
• Override the PrintTicket() method in child classes but this will duplicate the printing logic which is violation of DRY(Do not repeat yourself) principle.

This depends on the way the PrintTicket is implemented. If the method needs to consider information from subclasses, it needs to provide a way for them to supply additional information.

Moreover, you can override without repeating your code too. For example, if you call your base class method from the implementation, you avoid repetition:

class ConcessionTicket : Ticket {
    public override string PrintTicket() {
        return $"{base.PrintTicket()} (concession)"
    }
}

How can I satisfy above business requirement without violating the Open/Closed principle?

Template Method pattern supplies a third option: implement PrintTicket in the base class, and rely on derived classes to supply additional details as needed.

Here is an example using your class hierarchy:

abstract class Ticket {
    public string Name {get;}
    public string Train {get;}
    protected virtual string AdditionalDetails() {
        return "";
    }
    public string PrintTicket() {
        return $"{Name} : {Train}{AdditionalDetails()}";
    }
}

class RegularTicket : Ticket {
    ... // Uses the default implementation of AdditionalDetails()
}


class ConcessionTicket : Ticket {
    protected override string AdditionalDetails() {
        return " (concession)";
    }
}

What are the criteria to consider class is closed for modification?

It’s not the class that should be closed to modification, but rather the interface of that class (I mean “interface” in the broad sense, i.e. a collection of methods and properties and their behavior, not the language construct). The class hides its implementation, so the owners of the class can modify it at any time, as long as its externally visible behavior remains unchanged.

Is it after initial implementation of the class or may be after first deployment in production or may be something else?

The interface of the class needs to remain closed after the first time that you publish it for external use. Internal-use classes remain open to refactoring forever, because you can find and fix all its usages.

Apart from the simplest cases, it is not practical to expect that your class hierarchy is going to cover all possible usage scenarios after a certain number of refactoring iterations. Additional requirements calling for entirely new methods in the base class come up regularly, so your class stays open to modifications by you forever.

Let’s start with simple line of Open-closed principle – "Open for extension Closed for modification".
Consider your problem I would design the classes so that base Ticket would be responsible for printing a common Ticket information and other Ticket types would be responsible for printing their own format on top of base printing.

abstract class Ticket
{
    public string Print()
    {
        // Print base properties here. and return 
        return PrintBasicInfo() + AddionalPrint();
    }

    protected abstract string AddionalPrint();

    private string PrintBasicInfo()
    {
        // print base ticket info
    }
}

class ConcessionTicket : Ticket
{
    protected override string AddionalPrint()
    {
        // Specific to Conecession ticket printing
    }
}

class RegularTicket : Ticket
{
    protected override string AddionalPrint()
    {
        // Specific to Regular ticket printing
    }
}

This way you’re enforcing any new type of ticket that would be introduced in the system will implement it’s own printing feature on top of basic ticket printing information.
Base Ticket is now closed for modification but open for extension by providing additional method for it’s derived types.

The problem is not that you are violating the Open/Closed principle, the problem is that you are violating the Single Responsibility Principle.

This is literally a school book example of an SRP problem, or as wikipedia states:

As an example, consider a module that compiles and prints a report. Imagine such a module can be changed for two reasons. First, the content of the report could change. Second, the format of the report could change. These two things change for very different causes; one substantive, and one cosmetic. The single responsibility principle says that these two aspects of the problem are really two separate responsibilities, and should therefore be in separate classes or modules. It would be a bad design to couple two things that change for different reasons at different times.

The different Ticket classes may change because you add new information to them. The printing of the ticket may change because you need a new layout, hence you should have a TicketPrinter class that accepts tickets as input.

Your Ticket classes can then implement different interfaces to provide different types data or provide some kind of data template.

Modify the PrintTicket() method in Ticket abstract class. But this will violate Open-Closed principle.

This does not violate the open-closed principal. Code changes all the time, that’s why SOLID is important, so code remains maintainable, flexible, changeable. There’s nothing wrong with changing code.

OCP is more about external classes not being able to tamper with a class’s intended functionality.

For instance, I have a class A that takes a String in the constructor, and verifies this String has a certain format by calling a method to verify the string. This verifying method also needs to be usable by clients of A, so in the naive implementation, it’s made public.

Now I can ‘break’ this class by inheriting from it and overriding the verifying method to just always return true. Due to polymorphism, I can still use my subclass anywhere where I can use A, possibly creating unwanted behaviour in my program.

This seems like quiet a malicious thing to do, but in a more complex code base, it might be done as an ‘honest mistake’. To conform to OCP, class A should be designed in such a way that making this mistake is not possible.

I could do this by, for instance, making the verifying method final.

Inheritance isn’t the only mechanism that can be used to fulfil the requirements of OCP, and in most cases I would argue that it is rarely the best — there are usually better ways, as long as you plan ahead a little to consider the kind of changes you’re likely to need.

In this instance, I would argue that if you expect to get frequent format changes to your ticket printing system (and that seems like a fairly safe bet to me), using a templating system would make a lot of sense. Now, we don’t need to touch the code at all: all we have to do to fulfil this request is change a template (as long as your template system can access all the data that’s required, anyway).

In reality, a system can never be entirely closed against modification, and even getting it close would require a huge amount of wasted effort, so you have to make judgment calls over what kind of changes are likely, and structure your code accordingly.