Data Oriented Programming in Java 21, Solving the Countdown game - JEP Cafe #22

In this JEP Café, we will use a lot of the new features added to the JDK in the context of the Data Oriented Programming support: Records, Sealed Types, Pattern Matching, Unnamed Patterns, Exhaustive Switches, as well as the most recent additions to the Stream API: mapMulti(), and Gatherers (a preview feature of the JDK 22). As a bonus, you will see an anonymous class used as a non-denotable type in action, to create a small wrapper on a mutable content. All this is used to solve a funny problem: the Countdown game, where you need to reach a target with a set of randomly drawn numbers, and the four arithmetic operations. Let us check if all this can take us to an efficient, and readable solution.

⎯⎯⎯⎯⎯⎯ Chapters ⎯⎯⎯⎯⎯⎯

0:00 Do you feel like a coffee break?
0:11 Introducing the Countdown game and its rules
03:01 Introducing the Haskell algorithm translated to Java
04:44 Computing all the permutations of all the subsets
07:20 Splitting the sets to apply the four operations
08:53 Filtering the invalid operations
09:55 Computing the final result and the string of characters
10:21 Using Data Oriented Programming to model the data
12:05 Wrapping up this algorithm
14:49 Writing a bottom up algorithm leveraging Data Oriented Programming
17:32 Using Optional to discard invalid mapping operations
21:37 Implementing resolve() with an exhaustive switch and pattern matching
23:00 Using a Gatherer to inserting an element at a specific place in a sorted stream
27:18 Using anonymous classes and non-denotable types
28:23 Using the Gatherer in the final process
29:00 Making the algorithm efficient
30:59 Writing the final algorithm using Records
31:50 Comparing the two algorithms
32:33 That's it for today, talk to you soon!

⎯⎯⎯⎯⎯⎯ Resources ⎯⎯⎯⎯⎯⎯

Some amazing Countdown episodes!
More resources on Data Oriented Programming and other subjects covered in this JEP Café

Tags: #Java #Java21 #OpenJDK #JDK #JDK21 #JEPCafe #insidejava

[giovanninabil]

Finally, Java is flexing on its multi-paradigm design

[emileastih]

Great video as always José, with some truly insightful content! However, I personally miss the old format where you would code live and explain your thought process as you go.
It helped me understand not just the 'what, ' but also the 'why' behind each line of code.
Looking forward to more great content!

[mirogula]

Very good, in depth explanation. Thank you.

I wrote similar algorithm for solving countdown numbers game long time ago, when I was still at university. It was challenging, but fun problem to solve. I spent couple of weeks on it, on and off, and in the end I come up with quite elegant general solution, although I don’t know if it was as efficient as the solution described in your video.

[AleksandarT]

Great ideas but complex code in general :)

[Ambusher]

We should have another one where Java's support of Functional / Data-oriented programming paradigms is investigated. It can also detail which Java libraries are built on which paradigm, etc

[anyone-di]

Hi, guys! Why there are no more any new videos of Java Cafe? I'm so waiting for them :)

[TehBrian]

Is the DIV case at 21:24 supposed to have `resolve(left) % resolve(right)` instead of `resolve(right) % resolve(right)`?

[jay_sensz]

28:23 I don't think you really need that initializer block, you might as well just assign the value in the field declaration directly, e.g. `Element previous = toBeInserted;`

[danwhitfield-wkbf]

Where can I get that Java coffee cup? I can’t seem to find it anywhere!?

[nigelt]

Where can we get that Java coffee mug ❤😊

[dhankumarbajoliya]

when @Functional interface contains only one abstract method then why comparator interface contains two abstract methods

[baronvonherzenberger]

Why does everything need to be a gimmick? What is wrong with just presenting the code and functionality? Adult programmers don't need gimmicks like coffee intro and backstory. Just get to the point please.

[ChannelCheesecake]

😂 “Data” oriented programming. It’s just functional programming that is old as shit

[ampcgmail]

Why compareElement(e1, e2) method has to rely on the premise e2 >= e1 in their implementation Integer.compareTo(e2, e1) ? Integer.compareTo(e1, e2) doesn't works