Jump to content

  • Log In with Google      Sign In   
  • Create Account

Interested in a FREE copy of HTML5 game maker Construct 2?

We'll be giving away three Personal Edition licences in next Tuesday's GDNet Direct email newsletter!

Sign up from the right-hand sidebar on our homepage and read Tuesday's newsletter for details!


We're also offering banner ads on our site from just $5! 1. Details HERE. 2. GDNet+ Subscriptions HERE. 3. Ad upload HERE.


Relearning how we program in Java: The later stages of program development


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
1 reply to this topic

#1 tom_mai78101   Members   -  Reputation: 575

Like
0Likes
Like

Posted 23 October 2012 - 12:25 PM

Suppose you're into the later stages of a Java project's development, chances are you probably will see at least 1 Java source containing an object that you probably wish it should either be an abstract class or an interface. If you're the only developer working on this project, you probably will see plenty of places where you might wish they are "abstractable" or "interfaceable", and probably started thinking to yourself that one day, you'll return to this area and start refactoring the codes or do a small rewrite. (Not full-scale rewrite or anything.) This is the area I'm interested in, a thing that just kicks in when you know that something doesn't feel right.

You'll probably hear a few of these rules during your programming sessions (They are probably either not accurate enough, or they are completely false in every way):
  • If you're seeing common code structures, they are to be restructured into a method. If you're seeing common methods, they are to be restructured into a method placed in a superclass or something static.
  • If you're seeing objects with a common ancestor (or there appears to have an is-a relationship), they are to be restructured into an abstract class.
  • If you're seeing common methods declared in a few classes in general, those methods are to be restructured into interfaces.
Sometimes, a Java programmer would start off with an interface and expand from there. Sometimes, another dev guy would start from an abstract class, and evaluate to see if it needed interfaces or continue extending the class down. Other times, devs would start off with a basic class extending from Object, and continue to add/pile up additional code until it's ready to be refactored.

What is in your mind telling you and your guts that you should start using the rules mentioned (if any of them are correct)? And what makes you want to segregate parts of certain abstract classes into interfaces that follows the rules (if any are correct)? And while expanding objects (by adding more code to a regular class), when should you trust your instincts that you need to split them into either abstract classes, abstract methods, interfaces, or make them a subclass of another class?

To me, learning and understanding this will allow me to plan ahead in the very-near future of a simple concept of idea implementation in mind.

For example, a flexible bicycle. To make it flexible, I could start off with different interfaces. Then, I could create an abstract class, Bicycle, that implements all of the methods I will 100% use from the interfaces I will also 100% use. Finally, I could extend from the abstract class, and create a few more objects with a common design. Because I start off from creating Interfaces -> Abstract Class -> Subclass of the abstract class, my program design is crisp and clear. If I start off from creating a class, I might end up tiring myself in the future if I were to add a few more similar objects with similar methods.

Sponsor:

#2 EpicWally   Members   -  Reputation: 282

Like
0Likes
Like

Posted 23 October 2012 - 02:35 PM

I'm not an expert programmer, but I have been reading a lot of books recently on best practices and agile programming, so feel free to take this advice with a large grain of salt.

You cannot plan for everything up front. Think of a city. It starts as a town, with a small town square, and just a couple necessary buildings. Then it expands, adding in a few stores. Eventually some buildings get ripped down and replaced with strip malls. Roads eventually get ripped out and replaced with highways. etc.Coding necessarily has to follow the same progression, as it is a huge task to plan everything up front. Besides, you will always run into roadblocks along the way, and you will need to be adaptable.

My opinion is that your best bet, is to cover as much of your project as possible with unit tests. This way, you can fearlessly try refactoring your code into cleaner structures every time your program needs to evolve. You will know you haven't broken anything if all the tests still pass.

Most of what's above is shamelessly stolen from "Clean Code - A Handbook of Agile Software Craftsmanship". I simply copied/paraphrased to make the relevant points easily accessible. I highly recommend this book.




Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS