Daryl Teo's Blog

Learning C#: Early Impressions

Dec 21, 2008 Articles

For the last couple of weeks, I have been spending time learning C# (the mongrel offspring of Java and C++, as I like to describe it). So far, I haven’t had much trouble, thanks to my strong background in Java, and I think I have done enough in a week to be able to compare the two languages and talk about the parts I like and the parts I don’t. Perhaps these will grow on me in the future, or it may not.

This is not a rant about ASP.NET, which I have also been learning at the same time. That will come in time.

Namespaces vs Packages

At the moment I am leaning towards Packages. Packages in Java are heavily enforced, by making sure the .java is placed in the right spot within the file structure of your project. Namespaces, however, allow the “assembly” to be placed pretty much anywhere in the project. Although it is possible to follow a strict convention like Packages, it is really up to the developer.

Strangely enough, C# does not use C/C++’s namespace system (where a namespace can be used anywhere in the code) by enforcing all namespace declarations to the top (a la Java Packages), having them apply to all classes in that particular assembly. In the end, however, minimal adjustments were needed for me to get used to the C# Namespace system.

Get/Set vs Accessors/Mutators

I have mixed feelings about C# Properties. On the one hand it provides straightforward access to properties of an object, while encapsulating object members properly. On the other hand, I feel it obfuscates the code to a small extent. Let us compare two snippets of code.

C#

Java

In C#, the “Action” is not clear to me. Yes, we know it is assigning some boolean value to “true”. But if we look at the Java version, the method clearly states the intention. Furthermore, while C# Properties provides ways to implement read-only attributes, this is often unclear without the use of documentation. With Java, it is easy to see what operations are allowed with a particular class. This makes code easier to understand for me.

Another thing I have mixed feelings about is Interface Properties. While the intentions are sound, it breaks my concept of an Interface being a definition of functionality (as opposed to Superclasses, which should be a common set of properties). I suspect this is a “many tools for a single problem” syndrome that is often found in C++.

I may get used to Properties eventually, but it is not impossible to implement member access using methods instead of Properties.

C# Protected vs Java Protected

Here I have to say C# is the winner. Java’s Protected modifier was often used to provide member access to subclasses. However, it broke encapsulation by allowing access to other classes in the same package. C# follows C++’s version of Protected, by only allowing member access within the class, and derived classes only.

Subclass Declarations

Unfortunately, this aspect of the C# language is also inherited from C++ (pun intended). Java’s declarations are much clearer, and you can clearly see what are classes and what are interfaces, despite the more lengthly declaration. For example

C#

Java

Here we can see a problem. In C# there is no way to identify which is the superclass (or even if it does derive from one), and which are the interfaces, since there is no clear semantic. The “suggested” solution to this is to enforce a naming convention for Interfaces.

Again, a small problem, and I suspect C++ fans will enjoy their version alot more.

Conclusion

At this point in time, if I had to choose between the two, I would definitely go with Java. That’s not saying Java is better (the JVM is quite horrible, actually), but rather I prefer the language. As I like to say, “There is no way to go wrong in Java. If there is something wrong, its almost definitely your fault”. C# suffers the same “problems*” as C++; for a single problem it has many varied solutions. For example, I have never been a supporter of operator overloading. If someone overloads the + operator, how can you be sure that it adds two things together? You end up having to fight the language, by having to look through documentation all the time to figure out what it does.

Perhaps I will get used to it, perhaps not.

*I say “problems”, but they’re just my personal opinion.

Tags: C#, Comparison, Java, Programming