So whatever happened to the whole PGDCE thing?

[pitfiend]

...One thing I would definitely do immediately (if it was up to me) is completely scrap all use of that nightmarish fake-generics "template" library. Talk about include-file hell! There is no logical reason for anyone to be using anything other than Generics.Collections for their container classes these days, whether on Delphi or FPC/Lazarus. (The FPC version is actually better, BTW.)

This was a topic of very deep debate, we decided to drop generics, because there are differences between Delphi and FPC on how they tackle them. So, as we wanted to provide an engine for everybody, there are some sacrifices to do and some requirements to meet.

[Akira13]

This was a topic of very deep debate, we decided to drop generics, because there are differences between Delphi and FPC on how they tackle them. So, as we wanted to provide an engine for everybody, there are some sacrifices to do and some requirements to meet.

FPC is 100% compatible with Delphi-style generics when you use the {$mode Delphi} define though, which you guys did in fact have set for FPC in "PGDCE.inc". So you could have just used the Delphi syntax and not had any issues on either compiler.

[pitfiend]

FPC is 100% compatible with Delphi-style generics when you use the {$mode Delphi} define though, which you guys did in fact have set for FPC in "PGDCE.inc". So you could have just used the Delphi syntax and not had any issues on either compiler.

we decided to drop generics because we planned to bring the engine to developers on older delphi with no generics.

[JC_]

People are talking here about APIs.. it's unnecessary debate If is somebody making 3D game so it's clear that he will use Unity, Unreal Engine or something like that.
Using something else is 100% nonsense. For 2D no matter if you have OpenGL 1 or 4.5 and it's question about framework. Simply use something which have living community like SDL or SFML
because all Pascal engines/frameworks are dead, it's one man show, looks prehistoric or they are super slow - anyway no one is using them.

If there was anything usable so it was ZenGL, but today it's useless also, unfortunately.

I think we should more focus on making tutorials about various game mechanics like:

• managing game states
• working/managing with in-game entities
• path finding algorithms
• AI
• procedural assets creation
• physics
• and so on

This is good idea, but instead of tutorials we can create a community which share and improve code and where people can learn something.
The code should be optimized and easy to use. It has some perspective. I saw interesting things around like Coolsprite (something like Spine 2D) etc, also everyone need particle engine,
sound manager..

[SilverWarior]

If is somebody making 3D game so it's clear that he will use Unity, Unreal Engine or something like that.
Using something else is 100% nonsense.

Why do you say using something else is nonsense?
Tell me what would prefer more? Using Unity or Unreal Engine and with it being forced to do most of the programming in some other programming language than Pascal? Or seeing that comparable pascal based game engine is being built so you could do all of the programming in Pascal?
Well the idea of PGDCE is to make the later option a reality. But off course such task isn't trivial and I fear many of PGD members are afraid and insecure in their knowledge in order to participate in such project.

For 2D no matter if you have OpenGL 1 or 4.5 and it's question about framework.

I strongly disagree with you on this. Even when rendering 2D graphics you could greatly benefit from features available in newer API's.

Simply use something which have living community like SDL or SFML because all Pascal engines/frameworks are dead, it's one man show, looks prehistoric or they are super slow - anyway no one is using them.

Why do you think all Pascal engines/frameworks are dead? Because there weren't enough interest in using them. And why weren't enough interest from people in using them? One of the reasons is people always saying of how much they suck instead of trying to help improving them. Pretty much like you now.
The hard truth is that if we don't get together and start working on them nobody will.

This is good idea, but instead of tutorials we can create a community which share and improve code and where people can learn something.
The code should be optimized and easy to use.

Why should the posted code be optimized? So that you can only copy and paste it into your project? This way you won't learn much especially because heavily optimized code is often much harder to understand.
I would recommend that instead of worrying about the posted code being optimized we provide as many possible approaches in solving specific problems. This way we allow people to chose the approach which suites them the most. And that largely depends on their coding style and the rest of the project code.
Also once people truly understand some gane concepts they will be able to optimize the code on their own. And yes optimizing code for some game concept also heavily depends on the rest of the project code.

And no I'm not one of those people who say "Best optimization is no optimization at all". I myself like to have nicely optimized code but I try to avoid doing any premature optimizations but I still sometimes get carried away with premature optimization. Well nobody is perfect.

[turrican]

Hi,

Most of the time Object Pascal compiled code is near fast as C++ code. I develop applications and games with Delphi/FPC (production enviorment) and never have perfomance problems. Every language developer can do good programs and bad programs. The important thing is the developer and the spirit.

I use ZenGL to make games for Android, Windows and Linux, and all the games I developed runs at 60fps at most systems with FULLHD graphics, more FPS than crap Android Java SDK based games. Also I'm trying to port Ñuño's AllegroPAS to Android.

I think the developer is the whole think and programming language is a secondary stuff.

[JC_]

Why do you say using something else is nonsense?

If you're making a real 3D game so you need complex and standardized tools with some continuity and with good visual/graphics possibilities (in case that you are not John Carmack).
It's not only about people but also about time and money. There are some good non-commercial 3D engines? Or exists more commercial engines which can compete Unity, UE, CryEngine?
I like games like Legend of Grimrock with own 3D engine which have benefits from optimalization etc., but this is another case. They have team of people making engine directly for the game
not not kind of universal thing. Secondly, I have not seen such game with own 3D engine done in Pascal or in Java etc. => if somebody want 100% performance logically he uses the language that will allow it.

I strongly disagree with you on this. Even when rendering 2D graphics you could greatly benefit from features available in newer API's.

Can you specify which benefits and features? OpenGL have shader support since 2004, which was maybe most important thing for 2D graphics rendering?

Why do you think all Pascal engines/frameworks are dead? Because there weren't enough interest in using them. And why weren't enough interest from people in using them? One of the reasons is people always saying of how much they suck instead of trying to help improving them. Pretty much like you now.

If you need more real reason: 90% of them were not usable just some start-up projects. Working systematically is challenging. In second case, capable author(s) lost interest ( own life, no time, growing complexity, limitations of Pascal language..).
You can use same concept again and it will probably fail again. I do not know how you, but I preferred the solution which survived such natural selection and are universal like SDL or SFML..
This increases the chances to finish game in 1 person lifetime.

Why should the posted code be optimized? So that you can only copy and paste it into your project? This way you won't learn much especially because heavily optimized code is often much harder to understand.
I would recommend that instead of worrying about the posted code being optimized we provide as many possible approaches in solving specific problems. This way we allow people to chose the approach which suites them the most. And that largely depends on their coding style and the rest of the project code.

I do not understand much here, eg. ZenGL code is well optimized and very easy understandable and modifiable. On other hand there is ZenGL Particle editor from different author who did great job but it's pure chaos.
Chaos is not optimized in that sense

[turrican]

If there was anything usable so it was ZenGL, but today it's useless also, unfortunately.

You are wrong... Today is still usable and stable. But need developers to mantain it! I think we need to focus on this framework because is simple and very strong.

[Akira13]

we decided to drop generics because we planned to bring the engine to developers on older delphi with no generics.

I see. Personally I wouldn't have bothered with trying to support versions that old, but that's just me. I mean, if you can't afford a newer Delphi version, why wouldn't you just use FPC/Lazarus instead? I'd rather quit programming forever than have to use Delphi 7 in this day and age...

You are wrong... Today is still usable and stable. But need developers to mantain it! I think we need to focus on this framework because is simple and very strong.

It's an abandoned project (the developer seems to have just vanished) built ENTIRELY around outdated OpenGL functionality. I don't doubt you get 60 FPS in what are probably relatively simple 2D sprite games, but that's just your CPU brute-forcing the immediate-mode rendering. If you tried to scale ZenGL up to anything relatively complex you would hit a brick wall, trust me.

[turrican]

It's an abandoned project (the developer seems to have just vanished) built ENTIRELY around outdated OpenGL functionality. I don't doubt you get 60 FPS in what are probably relatively simple 2D sprite games, but that's just your CPU brute-forcing the immediate-mode rendering. If you tried to scale ZenGL up to anything relatively complex you would hit a brick wall, trust me.

Yes, simple 2D games are my focus now. No CPU usage at all (2%) with more than 800 sprites at screen. What modern framework do you recommend me to have the same functionity?