Phaser 4 Announcement

Published on 18th July 2019

Hi everyone - I want to share with you some exciting news that paves the way for the future of Phaser in 2019 and beyond. It’s with great pleasure I can announce that thanks to the continued support from the Facebook Gaming and the Instant Games teams, I’m able to start the development work required to build the next iteration of Phaser. This post will outline what’s planned and hopefully answer many of the questions you may have.

Phaser has been working together with Facebook Gaming for some time now. We initially started work together back when we jointly released the Instant Games plugin and associated tutorials for Phaser 3. The working relationship has been great and has allowed us to come up with the next evolution of Phaser together.

The continued support from Facebook directly allows for something I’ve been wanting to do with Phaser for a while now: Rewrite it in TypeScript.

Those of you who have been around Phaser for long enough will remember it was originally written in TypeScript, way back before TypeScript was even at a 1.0 release. We moved away from it due to an endless stream of bugs with the TypeScript compiler and tools. Quite frankly, we had adopted it far too early and got burnt badly in the process. Fast forward 5 years and the world is a very different place. TypeScript has matured beyond all recognition and the tooling around it now is best-in-class. What’s more, lots of you are now using it for your game development projects, too. Far more than ever before if our TypeScript defs stats are anything to go by.

Let me clear up a few important points now:

  1. This is not an API rewrite. Phaser 3 is comprised of hundreds of small modules already, so the work will be focused on porting those to TypeScript. Yes, I will absolutely use this process to improve the API as I go. There’s no way I will make any claims about “the API not changing” because I know it will. The core work, however, isn’t to rewrite or restructure, it’s to port.
  2. You will not need to use TypeScript in order to use Phaser. Phaser releases will still be published as ES5 bundles. If you code in ES5 JavaScript nothing about your workflow need change.
  3. The TypeScript version of Phaser will be Phaser 4.0. I really didn’t want to have to bump the number like this, because for Phaser the major number is part of the name, not the release version, but I feel it’s an important enough update that I have no choice but to mark it in this way. It will also mark a clean point of separation for users.
  4. Even with this continued support from Facebook, I still absolutely need your support via Patreon. All of my projections are based on a combination of resources and support from one source does not cancel out the others. It’s still vitally important to the project that Patreon continues.

Why TypeScript and not ES6?

Although I’ve been discussing this work with Facebook for months now, I hadn’t made a call on if the port should be to ES6 or TypeScript. Facebook had indicated they were open to either option and personally I had no strong leaning in either direction either. A few things tipped the scales in TypeScript’s favor, though. The first is that I’ve sunk a lot of resources into our current TypeScript definitions generator. While it works, it’s less than ideal and we are forever tweaking the output to make it more accurate for TypeScript developers. This is a problem beset by all projects that use tools to automate the creation of TypeScript defs and I see issues around it across many, many projects outside of Phaser. As more and more developers use TypeScript as their primary language, it will give them a perfect developer experience, while taking nothing away from JavaScript developers.

Recently a few bugs have slipped into the API that would have been solved by using strict types, which TypeScript will bring. Although most bugs are, and always will be, logic rather than type related - catching even just one type bug before it hits production is a win in my books.

Finally, this move will significantly help those new to the API. If they choose to use TypeScript they’ll get full and 100% accurate code insight and context awareness. This is a huge benefit and shouldn’t be under-estimated. It will help me, too, as I work on the API, speeding up my development workflow as much as anyone else's. I spend hours and hours inside of VS Code each day and I’ll finally be able to benefit from the massive number of tools it contains. Time saved is time gained.

I’m expecting the development of Phaser 4 to begin by the end of August 2019. First will come the release of Phaser 3.19. I will be building it in a new branch off of the main repo. 3.x issues will still be addressed, and new versions published. As I near completion of 4.0.0 this process will halt. Until then, nothing will change.

I know better than to give any kind of estimation for when Phaser 4 will be published. And honestly, I don’t actually know. I’ve made estimates, of course, but I’m not committing to a date, no matter how vague. As with all versions previously, you will be able to track progress openly via GitHub and my Dev Logs.

I’m really looking forward to this next stage. Phaser 3 had quite a difficult birth, although that was due to scope and expectations, not technology. This work allows me to pull it kicking and screaming into the modern web era and tighten-up ship as we go. It’s not about reinventing for the 4th time, it’s simply about super-charging and streamlining what we already have. And that is genuinely exciting.

Hi everyone - I want to share with you some exciting news that paves the way for the future of Phaser in 2019 and beyond. It’s with great pleasure I can announce that thanks to the continued support from the Facebook Gaming and the Instant Games teams, I’m able to start the development work required to build the next iteration of Phaser. This post will outline what’s planned and hopefully answer many of the questions you may have.

Phaser has been working together with Facebook Gaming for some time now. We initially started work together back when we jointly released the Instant Games plugin and associated tutorials for Phaser 3. The working relationship has been great and has allowed us to come up with the next evolution of Phaser together.

The continued support from Facebook directly allows for something I’ve been wanting to do with Phaser for a while now: Rewrite it in TypeScript.

Those of you who have been around Phaser for long enough will remember it was originally written in TypeScript, way back before TypeScript was even at a 1.0 release. We moved away from it due to an endless stream of bugs with the TypeScript compiler and tools. Quite frankly, we had adopted it far too early and got burnt badly in the process. Fast forward 5 years and the world is a very different place. TypeScript has matured beyond all recognition and the tooling around it now is best-in-class. What’s more, lots of you are now using it for your game development projects, too. Far more than ever before if our TypeScript defs stats are anything to go by.

Let me clear up a few important points now:

  1. This is not an API rewrite. Phaser 3 is comprised of hundreds of small modules already, so the work will be focused on porting those to TypeScript. Yes, I will absolutely use this process to improve the API as I go. There’s no way I will make any claims about “the API not changing” because I know it will. The core work, however, isn’t to rewrite or restructure, it’s to port.
  2. You will not need to use TypeScript in order to use Phaser. Phaser releases will still be published as ES5 bundles. If you code in ES5 JavaScript nothing about your workflow need change.
  3. The TypeScript version of Phaser will be Phaser 4.0. I really didn’t want to have to bump the number like this, because for Phaser the major number is part of the name, not the release version, but I feel it’s an important enough update that I have no choice but to mark it in this way. It will also mark a clean point of separation for users.
  4. Even with this continued support from Facebook, I still absolutely need your support via Patreon. All of my projections are based on a combination of resources and support from one source does not cancel out the others. It’s still vitally important to the project that Patreon continues.

Why TypeScript and not ES6?

Although I’ve been discussing this work with Facebook for months now, I hadn’t made a call on if the port should be to ES6 or TypeScript. Facebook had indicated they were open to either option and personally I had no strong leaning in either direction either. A few things tipped the scales in TypeScript’s favor, though. The first is that I’ve sunk a lot of resources into our current TypeScript definitions generator. While it works, it’s less than ideal and we are forever tweaking the output to make it more accurate for TypeScript developers. This is a problem beset by all projects that use tools to automate the creation of TypeScript defs and I see issues around it across many, many projects outside of Phaser. As more and more developers use TypeScript as their primary language, it will give them a perfect developer experience, while taking nothing away from JavaScript developers.

Recently a few bugs have slipped into the API that would have been solved by using strict types, which TypeScript will bring. Although most bugs are, and always will be, logic rather than type related - catching even just one type bug before it hits production is a win in my books.

Finally, this move will significantly help those new to the API. If they choose to use TypeScript they’ll get full and 100% accurate code insight and context awareness. This is a huge benefit and shouldn’t be under-estimated. It will help me, too, as I work on the API, speeding up my development workflow as much as anyone else's. I spend hours and hours inside of VS Code each day and I’ll finally be able to benefit from the massive number of tools it contains. Time saved is time gained.

I’m expecting the development of Phaser 4 to begin by the end of August 2019. First will come the release of Phaser 3.19. I will be building it in a new branch off of the main repo. 3.x issues will still be addressed, and new versions published. As I near completion of 4.0.0 this process will halt. Until then, nothing will change.

I know better than to give any kind of estimation for when Phaser 4 will be published. And honestly, I don’t actually know. I’ve made estimates, of course, but I’m not committing to a date, no matter how vague. As with all versions previously, you will be able to track progress openly via GitHub and my Dev Logs.

I’m really looking forward to this next stage. Phaser 3 had quite a difficult birth, although that was due to scope and expectations, not technology. This work allows me to pull it kicking and screaming into the modern web era and tighten-up ship as we go. It’s not about reinventing for the 4th time, it’s simply about super-charging and streamlining what we already have. And that is genuinely exciting.