Defold video channel

[selimanac]

We are the ordinary three man Indie. One client dev, one game/art designer, one server dev.
I choose Defold as a client dev because it is a (almost)bare-bones 2D game engine with simple but great components. You can simply focus on your game and your code. You have the control over the performance, libraries, shaders, render pipeline, UI… Native Extensions are big bonus. It is a huge advantage that your questions are answered(fast) by core Defold developer team and other talented developers. You can’t buy or easily find that!
But before I decided to use Defold for production I spent almost a year with it. Yes I’m hitting walls and get mad with Defold sometimes. But this is the part of the game dev progress and not just special to Defold.

I must say that %90 of Unity community is sucks. Newcomers, kids, hobbyists… Most of the answered questions and assets are junk too. And who wants to develop a game with premade assets anyway… If you want control, you should do it by yourself.

Defold or Unity or whatever! is not important. Choosing correct tools for the right kind of game is the most important thing in game development. It is related to experience. If you know what you are doing and the engine’s limitations well none of these are the real issue… Also solving problems by myself is a gift for me, It teachs a lot.

Advanced particle editor and shader tutorial please

[Pawel]

I know other devs are in similar situation as I am. They are focusing on finishing games and will publish more for the community after.

Yes, indeed. I’m strugling to complete a boss level for UDGJ#1 meanwhile releasing Defluid and Defork for my game and still have plans to share a lot of stuff when I find more time to clean and describe them

I tried making my game in both Godot and Unity and it was hard to modify assets to my needs as I am able to do in Defold. (Well, I also tried Game Maker, Game Salad and Corona, but don’t want to say anything about them) Also worth mentioning is that Unity is showed in 2005 and Godot since the beginning of 2014, so they are older than Defold.

And BTW, I’m very excited of what @Pkeod is planning to publish and waiting for it impatiently!

[Mitch]

Yes, it is true that Defold is a part of King which is a part of Activision Blizzard. But King and AB are companies in the business of making games, not game engines. Unity on the other hand is (well, and to sell Ads). We will never assign hundred of engineers/hundreds of thousands of man hours and/or tens of millions of dollars in making Defold the number one game engine for everyone.

My comments were on other people saying it’s unfeasible to for King to create more content, tutorials and documentation for their engine. And I still stand by that it is 100% easily possible for them IF they want to.

You’re moving the goal posts. It’s a very big difference between having more documentation and tutorials etc which could be easily done by either outsourcing it to smaller youtube content creators or having 1-2 employees work full time on creating content and requiring hundreds of engineers, 100,000+ hours and tens of millions of dollars. Those 2 things aren’t even in the same field.

Could King reallocate 1 employee to full time content creation for Defold, employ a new staff or pay for outside content creation? I would be shocked if they couldn’t if they wanted to. The issue is whether or not they want to do it, like I said initially.

While we want a wide variety of developers to use Defold it is mainly a tool for experienced developers

How on earth is the implementation of loot tables so intrinsically connected to a specific game engine?

If your target market is focused on ‘experienced developers’ then is it really necessary to make videos reading documentation or teaching github. I would say creating loot tables the ‘Defold way’ is ‘more intrinsically connected to a specific game engine’ than github basics are.

shall we just read the manual aloud (like we do for the addressing videos)

Do you know if they were commissioned by Godot and Unity or if he created them on his own?

From my understanding created on his own.

We should educate our users (through manuals and/or tutorials) about the following things in Defold:

• Defold uses Lua
• The only data structure in Lua is the Table
• The basic building block is the Game Object
• Game Objects are created using Factories
• You use Collision Objects to detect when two Game Objects collide

Yup and if Defold had a similar tutorial resource it would cover the Defold way of doing a project from start to finish. It would be an invaluable addition that would draw in more users some of which would be the ‘experienced developers’ you seek.

This is where we should spend our time and maybe we need to spend more time doing this?

Definitely. I’m surprised to see so much kickback for suggesting that Defold needs more fleshed out documented quality content after a Defold employee asked for suggestions for video content.

[britzl]

My comments were on other people saying it’s unfeasible to for King to create more content, tutorials and documentation for their engine. And I still stand by that it is 100% easily possible for them IF they want to.

You’re moving the goal posts. It’s a very big difference between having more documentation and tutorials etc which could be easily done by either outsourcing it to smaller youtube content creators or having 1-2 employees work full time on creating content and requiring hundreds of engineers, 100,000+ hours and tens of millions of dollars. Those 2 things aren’t even in the same field.

What I’m trying to say is that King is not developing Defold to make money out of it. We’re quite happy with the current structure and we have no plans to compete with for instance Unity and the hundreds of developers they have working on that product.

Could King reallocate 1 employee to full time content creation for Defold, employ a new staff or pay for outside content creation? I would be shocked if they couldn’t if they wanted to. The issue is whether or not they want to do it, like I said initially.

We do have one technical writer assigned to the Defold team to create content for our Learn section. He’s currently on parental leave and there is no plan to replace him while he is gone. We could potentially pay for outside content creation (and we have to some extent in the past) but there are no plans to make anything big in that regard during 2019.

I’m surprised to see so much kickback for suggesting that Defold needs more fleshed out documented quality content after a Defold employee asked for suggestions for video content.

There’s a big difference between creating a 24 hour video tutorial covering the creation of a game from start to end and the creation of specific video tutorial content. We will continue to create short videos explaining specific areas and I believe that is what @Oleg_The_Evangelist was asking for.

[britzl]

PS If you do want to watch a large Defold game get created, hour by hour, then tune in to the Fates of Ort Twitch channel: https://www.twitch.tv/topbraj/videos

[Oleg_The_Evangelist]

I love this thread! A sincere thanks to each one of you who took their time and energy to reply here.

Now let me put my devil’s advocate hat here and actually back @Mitch, as he has a point.

Many Defold users (including the Defold team who also use Defold, prooflink) kinds of got used to what Defold is and how to use it. How to embrace the features and how to umm… skip/workaround/not focus on whatever Defold is not excellent at. And then we sometimes make everything harder for the newer Defold users, who don’t yet have the battle scars required to embrace Defold. Any many of them perhaps will drop Defold before earning that scars.

I personally did huge efforts to document all the onboardings where people were able to successfully befriend Defold and pass the steep learning curve. The problem was… there was little in common except for “I looked into the examples by @britzl, @benjames171, @sven @younameit and learned how to do things”. I haven’t figured out how to translate it into a successful onboarding video/manual/tutorial yet… But all ideas are welcome!

I am trying to say, the feedback focused on what Defold is not good at is also useful. Perhaps, we don’t think we are in a position to act immediately upon this particular case. But we did react fast each time @connor.halford brought in sets of frustrations, or when @marat.giliazov did it recently, because we knew how to.

Also this is the reason we are so open on how and why we prioritise feature requests and bugreports.

Finally, this post is a friendly reminder to myself to be open to all feedback, and value it even perhaps if there are not immediate action points at the moment.

[benjames171]

As a seasoned gamedev, I found Defold relatively easy to pick up and use. There were one or two novel concepts to get my head around at first such as collections and the way collision detection works. I’d looked at other engines before landing on Defold - Unity seemed like overkill for the type of game I wanted to make, too heavyweight and I’d never liked the way Unity put their branding on your game, at least with the free version. Godot was interesting but didn’t feel like a well engineered product the way Defold is. I really like the lightweight, minimalist approach to the design of Defold though it’s still feature-rich in my opinion, as @Oleg_The_Evangelist mentioned, there always seems to be a way of doing stuff even if it’s not quite as you’d initially imagined.

In terms of assets I think we need to be careful. Whilst having a large pool of pre-made assets might seem like a good idea, I feel what’s happened with Unity is a cautionary tale - that is, the continous output of games lashed together using stuff from the asset store. There is a temptation for new gamedevs to cut corners and not really learn how to do things properly. Things like memory and performance optimisation are not thought of at all, though granted these might not be the first prority for a newbie gamedev.

I learn best by looking at example code, which was the motivation behind releasing my own code for people to look at.

Finally, I love where we’re at with Defold right now. It feels like a secret weapon in my pocket and the Defold team’s continued dedication to keeping the engine compact and performant gives me great confidence in continuing to use it into the future.

[sergey.lerg]

It’s a good conversation going on here. I too think there are too few tutorials and examples for Defold, that’s why I am constantly joining conferences with talks on Defold and release my examples on GitHub. I have 3D RTS StarClick, 3D First Person View scene, Tiled platformer and a variety of native extensions and libs.
FPV scene is not currently on github, because I want to make it a youtube tutorial series.

Recently I’ve started learning Unity to see what’s the difference. Quickly I was put down on Earth - Asset Store is really difficult to find something useful. I couldn’t find a decent touch manager, so I coded one just like in Defold. With instacing of gameobjects in Unity everything feels very Defold like, but Unity has superior runtime features - creating meshes, changing materials, removing/adding components - I miss such stuff in Defold a lot. Plus proper 3D concepts - accessing vertex streams, working with normals and tangent, proper model loading.

Then again, I feel how Unity is so much bloated with stuff and it’s so easy to create laggy game that will eat all the resources. Sprites are much easier to create in Defold too.

Defold is relatively new and does not inherit design flaws like Unity. So I hope it will eventually grow into equally powerful engine while preserving it’s lightweightness and ease of use.

[Oleg_The_Evangelist]

I also think a thread like this, especially if coupled with a video lecture and a text tutorial is kind of a tool/process to ease the steep learning curve for Defold. The problem is, we have only one. And still somewhat unfinished. And even that required endless man-hours on our side.

Can this be outsourced? Well, theoretically. Practically we’d have to write lots of code/art/project conventions first. Which may be a useful exercise to do for Defold team, but it takes attention from other things, like fixing bugs or adding engine features.

Here-we-go. I’ve made another circle of what-ifs and should-woulds. Now that I’ve discovered a rabbit hole, I won’t jump into it Instead, I’ll remind about the teaser Fridays thread. We have an experimentation day today. So perhaps somebody decides to share their cool stuff today.

[Pawel]

Well, the good examples of how great our community is, are the latest extensions, tutorials, code samples and assets released I think recently we added such a vast amount of them, that it’s noticeable so, I think we should thank them again, @selimanac, @Insality, @sergey.lerg, @AGulev, @sven, @jhonny.goransson, @britzl, @benjames171 and generally all of you great job!

[gianmichele]

I have to say that probably the best training material are the one produced from real world scenarios, or in-house projects.

Take for example Blender open movies. Not only they are nice showcase products, but they are build to help the development of the software and produce training material.

I guess some of the latest Unity demo work in that direction too (Megacity).

[Oleg_The_Evangelist]

I will try to make videos with everybody who has shipped at UDGJ1, so they reflect on what why and how they did, while the memories are fresh.

Hopefully it could be a good reference on how to architect games in Defold.

My current plan is to focus on the following questions:

• What was your approach to design this game?
  • pen and paper prototype or straight to the code?
• How did you setup the project?
  • Did you start with empty project and then grew it up or did you add any favourite libraries first?
• How is your project structured technically?
  • MVC, some controller script? Everything in update? Some clever approach?
• How did you do the GUI/HUD?
  • Defold gui? Own gui?
  • Is GUI code/gos separate from the game logic?
• This is a jam project, so you did not have much time to do it right, so what were the shortcuts that you had to make?
  • How would you improve the game if you had more time?
• What was the hardest technical challenge or just a tricky thing to implement?
• Did you think Defold was missing a particular feature when making the game?
• Did you log any bugs/feature requests while making the game?

I am all open for the input on how do I make these videos more relevant to you all here. Don’t be shy in helping me do better videos =]

[Oleg_The_Evangelist]

Did the first recording today with @britzl. I have mixed feelings. Will try to ship it next week and then sit down with @sven and @jhonny.goransson to do the same.

In the meantime I’ve shipped few letsplay videos from the UDGJ1. Check them out:

[Oleg_The_Evangelist]

We tried a new concept for Defold tutorials.

I still believe that a step-by-step tutorials shall rather be done in text on a project, that was designed as a learning material with gradually increasing complexity. But there’s a huge demand for something that helps to walk the first mile with Defold. So maybe this is helpful enough?

1. Under 15 minutes, so the video is actually watchable;
2. The game in the tutorial is immediately available as a source code;
3. The game was made for a game jam, not as a learning material. Meaning the code may be harder to read, but it makes more sense for real life.

As always, all your feedback is appreciated.
@sven and @jhonny.goransson are next in line to be recorded.

[Oleg_The_Evangelist]

We’ve recorded a video with Sven now. It got fairly complex and almost 30 minutes long. I’ll cut at least 30% off and publish it in a week or so (next week is an internal King event).

Numbers for the video with Björn look very well. So perhaps we’re doing something right. But we can do better, so all feedback is appreciated.

[Pawel]

Video with Björn is really great, explanation for a code that is open-source is really great and should help users pretty well

[Oleg_The_Evangelist]

now talking to Sven about his game. Had to split into 2 parts for clarity and watchability. Will try to ship the second part of Sven tomorrow Then next week we all are at an event, so the promised video with Jhonny and LD44 recording will follow in about 2 weeks.

[Oleg_The_Evangelist]

more rocket science from Sven:

We’re sitting down with Jhonny next week to discuss his take on doing a game with Defold. Now that you have seen two takes - a lighter more generic one with Björn and a more rocket-sciency one with Sven - could anybody please reflect on what does not work in these videos?

Perhaps I/we should work better on a structure (i.e. architecture–>coding styles–>walk through one cool feature in the game–>bottom line) rather than doing what makes sense?

[Oleg_The_Evangelist]

Jhonny and his shady, ahem, shading technique featuring custom materials and some clever engineering.

Also I have a backlog of videos to come soon:

1. LD44 letsplays (those of you on twitter might have seen a teaser)
2. Making of Pet Rescue Puzzle Saga diaries
3. More vertical videos with Sven clap

[Pawel]

Making of Pet Rescue Puzzle Saga diaries

Can’t wait to see them, especially Pet Rescue