DonutKing wrote:

Another forum tried this a while ago... google search for 'zybourne clock' 🤣
If not for the hilarious 'art', it's at least a great example of how NOT to run a collaborative project.

"only one person on the team had previous game development experience as a professional developer. He was banned from the project for criticizing it."

Welp! 🤣

Jade Falcon wrote:

if you need any testers let me know. I don't have a true dos system but I can test in dosbox. If I do get ahold of a dos system I'll let you know.

OK!

Jorpho wrote:

May I suggest that you have spectacularly unrealistic ideas about the difficulties involved here?

That's why you guys are here, to tell me what can and can't be done. I'm sure something can be accomplished, it's just a matter of what exactly.

Jepael wrote:

I can mostly only code and debug. I think I could do real mode code, algorithms, sound drivers (SB), music drivers (Adlib, OPL3, CMS) and video drivers (VGA), in C, Turbo Pascal and assembly. Well, also input drivers like keyboard, mouse, and joystick.

OK cool! If this gets going that would be great.

VileRancour wrote:

Heh, it could be fun, but free time isn't very abundant lately (and what little there is tends to be absorbed in other hobby stuff, heh). I'm all for new games for old PCs, just to show up those C64/Spectrum/etc. guys who still churn them out regularly for their respective systems. So if there's something with really cool potential I'd want to lend a hand for sure, but without even a concept or a target it's kinda hard to say.

Tell me about it. In my humble opinion it's about time something like this happened. I'm churning over concepts in my mind right now, maybe I'll post a few ideas 😀

badmojo wrote:

🤣

Excuse me?

ZanQuance wrote:

Or an escape the room game before a Vogon reads his poetry

Heh, that would be funny.

Beegle wrote:

Level Designer by trade, but work has taken me elsewhere in the last few years. I miss making maps/gameplay/missions.
If people are cranked up for this, I'd like to give a hand.

Well sure! I'd love to have you help if we can get this going.

@DonutKing: That was hilarious. I love how they sacked the only competent person because he knew how they were doing everything wrong 🤣 Well, there you go, they failed, and if they'd been that stupid as a startup trying to make money they would have been broke, and come back to that guy begging him to help them 🤣

So I'm thinking this will either be a game for:

A. 286/EGA/SB(1.0/2.0) [Real mode]
or
B. 386/VGA/SB (1.0-Pro 2) [Real or protected mode]

I'm leaning towards B. Either way, it's a scrolling 2D platformer like the ones from the early 90's. What's easiest to develop for? My instinct tells me real mode but I have no idea if that's correct. I guess it's debatable whether graphics are easier to draw in 16 or 256 colors.

As for game story concepts, I need ideas thrown into the mix! Here's a few substandard ideas I came up with:

1. Take a concept from something in pop culture, for instance (a la 8088MPH) Back to the Future. Or something else like Star Wars, Star Trek, LOTR... I'm not saying we do this; I'm not so fond of the idea myself, but it's an idea all the same. If we did it at all I'd prefer Star Wars.

2. Resurrect a character from some other game, for instance Keen (overdone, there are a billion new Keen games already, based off of the Galaxy engine of course, so not exactly new but whatever), Duke Nukem, etc. I don't like this idea either but here it is.

3. Come up with something entirely new. Here are some ideas for that which have been knocking about in my head:
A. You are a CIA agent during the Cold War. Your job is to hunt down the KGB's most villianous agent and intercept him before he causes damage of some sort or something (yes, I know, lame story, but it could be improved upon). I can see opportunities for innovative features here; especially variations on the traditional "world map" level layout, in which you utilize air and automobile travel to get where you need to be.
B. You follow the adventures of a guy who builds a spaceship with warp drive and uses it to travel through the galaxy, where he gets captured by aliens, etc. Basically Keen, except the main character is an adult.
C. Fantasy universe thing. You control a guy who has to save the world from an evil sorcerer or something, and there's magic and dragons and a whole bunch of other fantasy stuff. Blah.

That's the best I can come up with right now. If I get any epiphanies I'll post them. Come on, guys, surely we have enough collective brainpower to invent a novel, innovative story line and gameplay, right?

OK, post your criticisms/corrections/etc. below. They're greatly needed.

probably something more original

like a elf girl protagonist with switchable friends that have different abilities and a soft rpg system to it all and backtracking and maps and a plot involving human conspiracy and witchcraft and subliminal environmentalist themes. also animated cutscenes and 6 different endings and make the game runnable off floppies

hopefully that doesnt sound derivative!!!

everyone has ideas, what matters is execution. but i got a feeling this thread will be stuck in the ideas/concept phase anyhow

oh and please no qbasic. i would probably suggest some form of c with djgpp - what Quake was programmed on and has its own extender, and supports Allegro which could be done for (slow) prototyping or ensuring modern-ish portability

Make it a strategy/simulation game: Abandonware Tycoon!

Your task is to hunt the internet for crappy cracked copies of old DOS games. The ultimate goal is to monopolize the adoration of the nostalgic masses, and to maximize your e-peen. However, you face fierce competition from *real* software preservationists! You must snatch up all available copies, build your reputation by offering them up for download, and quash your opponents by stealing popularity. Of course, to do this you'll have to keep as public as possible while still dodging anti-piracy action groups.

As you collect approval points, you'll be able to recruit slaves who do the work for you, and maximize your scores by appropriating their work and crediting yourself. You'll have to capitalize on quantity over quality, because the latter only slows you down. Once you get the coveted BEARD BONUS, the road to victory is assured: your public talks at PeenCon will defeat the copyright enforcers, your army of twitter lackeys will be harvesting free social capital for you, and you'll fully legitimize your warez site by giving it a Pinstagrumblr makeover and convincing the weekend nostalgics that your shit is gold. You have become... ABANDONWARE TYCOON!

😉

oh and please no qbasic. i would probably suggest some form of c with djgpp - what Quake was programmed on and has its own extender, and supports Allegro which could be done for (slow) prototyping or ensuring modern-ish portability

eh, DJGPP is inefficient, hasn't been optimized for DOS in ages (afaik) and pretty much precludes real mode...
From what I know, OpenWatcom is the way to go for protected mode. For pure real-mode, maybe a 'classic' borland compiler with performance-sensitive bits converted to assmebler as you go.

leileilol wrote:

hopefully that doesnt sound derivative!!!

No, that actually sounds pretty neat.

leileilol wrote:

everyone has ideas, what matters is execution. but i got a feeling this thread will be stuck in the ideas/concept phase anyhow

Well, I'll try to keep that from happening 😊

leileilol wrote:

oh and please no qbasic.

Absolutely. I pushed QB45 to its very limits with that engine I wrote, so I know that from experience 🤣

VileRancour wrote:

Make it a strategy/simulation game: Abandonware Tycoon!

Interesting, interesting. I like it. Question: what exactly is the "BEARD BONUS", anyhow?

VileRancour wrote:

eh, DJGPP is inefficient, hasn't been optimized for DOS in ages (afaik) and pretty much precludes real mode...
From what I know, OpenWatcom is the way to go for protected mode. For pure real-mode, maybe a 'classic' borland compiler with performance-sensitive bits converted to assmebler as you go.

I would tend to agree with this; leaning towards the Borland approach.

keenmaster486 wrote:

Question: what exactly is the "BEARD BONUS", anyhow?

an oblique reference that shall remain oblique, since I don't fancy being harvested by the aforementioned lackey army... and of course I totally wasn't referring to anybody in particular 😁

Yeah, wasn't a serious suggestion by any stretch, but you can go "meta" in other ways... I liked the idea of Retro Fever which was all about collecting vintage computers.

each machine needs a different one - find the right part for each computer before the roaming mutated lab rats get you

This made me laugh 🤣
As for your oblique reference, I guess I will remain in the dark 😉

leileilol wrote:

everyone has ideas, what matters is execution. but i got a feeling this thread will be stuck in the ideas/concept phase anyhow

That's the catch while doing a group effort to find "THE" idea. It takes forever and never has closure. And mostly everybody loves their own ideas.

My advice for keenmaster486 would be, instead of trying to find the idea, to determine instead a rough idea of the gameplay, and immediately do some prototypes.
The goal of doing prototypes so early would be to test different engines/languages/technology and try to settle on those.

So for the choices keenmaster486 proposed, here is my take :

Tech : B (386/VGA/SB) - because it allows more flexibility.
Story : 3 (something new) - will allow liberty on our side, and avoid copyright issues in the future.
Story : None - Too early. As mentioned earlier, a prototype would be more adviseable at this time.

You mentioned a side scroller. Try it out with the techier guys who introduced themselves earlier. See what comes out of it and if you manage to have something playable.
Note : "playable" here only needs to be a red cube that the player controls with arrows and press space to jump. In a test level. Nothing more. When you have that, we will go from there.

Beegle wrote:

You mentioned a side scroller. Try it out with the techier guys who introduced themselves earlier. See what comes out of it and if you manage to have something playable.
Note : "playable" here only needs to be a red cube that the player controls with arrows and press space to jump. In a test level. Nothing more. When you have that, we will go from there.

Okay done...what's next? 🤣 (sorry, I seriously couldn't resist)

I coded this as a joke years ago while I was learning how to program the Atari 2600 😁

YES I LOVE THE MYSTERIOUS RED SQUARE

Beegle, I think I completely agree with everything you said there, including your hardware & approach choices.

So, techier fellows, let's see what we can make here! 😀

So let's try to get some code whipped up, i.e. the foundation of the engine. I'm thinking real mode with the last DOS version of Borland C or Pascal, as VileR said earlier - does everyone think that's a good idea? [I will also say that FreeBasic is blazing fast in DOS, and very easy to use. I did some tests and it was some 20 times faster than QB45. Of course, that limits you to protected mode.]

Let's talk about the differences between real mode and protected mode. Afaik real mode can only address the first megabyte of memory - is that correct? What about EMS and XMS? Are those only available in protected mode? For our purposes I think I'd like to have more than just 600K or however much usable memory you actually typically get while in real mode. Is real mode faster than protected mode? Also, 16-bit or 32-bit protected mode? 16-bit protected mode works on a 286, right? It would be cool if the target platform is 386/VGA but it also runs on a 286/VGA.

How will we do graphics functions? Will we basically be writing a custom, optimized set of graphics functions for use with VGA? We'll have to implement something like adaptive tile refresh to make it scroll smoothly.

How should graphics be drawn? We could just have the graphics as BMP files (the tiles in sheets, the sprites individually) and whoever does the graphics can use their tool of choice. I also developed a custom sprite editor once that I really like 😀 I could modify it to load/save in whatever format is required. But the tiles, I think, ought to be large BMP files pasted together from various sources. I assume we're using 16x16 tiles.

ZanQuance: That's really cool! Now we recreate the idea of that for x86!

keenmaster486 wrote:

So let's try to get some code whipped up, i.e. the foundation of the engine. I'm thinking real mode with the last DOS version of Borland C or Pascal, as VileR said earlier - does everyone think that's a good idea? [I will also say that FreeBasic is blazing fast in DOS, and very easy to use. I did some tests and it was some 20 times faster than QB45. Of course, that limits you to protected mode.]

I suggest selecting tools that are freely available (open source or otherwise released to public), so anyone can easily build/develop/test the project.
Using protected mode limits the usage to 386 and above (I don't think you are talking about 16-bit protected mode of 286). Real mode works down to 8088, unless you use some 286 instructions to speed up some things. If a pmode environment is used, it should be really good and complete, so that calling real mode interrupts (DOS, video, mouse, etc) does not need to be handled manually.

keenmaster486 wrote:

Let's talk about the differences between real mode and protected mode. Afaik real mode can only address the first megabyte of memory - is that correct? What about EMS and XMS? Are those only available in protected mode? For our purposes I think I'd like to have more than just 600K or however much usable memory you actually typically get while in real mode. Is real mode faster than protected mode? Also, 16-bit or 32-bit protected mode? 16-bit protected mode works on a 286, right? It would be cool if the target platform is 386/VGA but it also runs on a 286/VGA.

Real mode allows access to 1MB of memory and with 286+ CPU there's almost 64KB of high memory area available as well. EMS is done with hardware on 8086/8088 so it is available in real mode, using 64KB window to access up to 32MB, but on 386+ the EMM386 driver uses protected mode to implement EMS with software, so everything runs under virtual 86 mode, not real mode. XMS is available under real mode, because on 286 and 386 the XMS driver switches the CPU to protected mode only when game asks to copy of data between conventional memory and XMS.

If EMM386 is installed so CPU is in virtual 86 mode, IO access is slower. Actual real mode is faster. 386 protected mode is faster as it allows native 32-bit code to run on 386, registers are 32-bit and you have more registers and better indexing modes available. 286 protected mode is not faster because it is still 16-bit code like in real mode, but of course allows access up to 16MB of physical memory.

A lot of good games run perfectly in real mode, utilizing XMS and/or EMS if available.

keenmaster486 wrote:

How will we do graphics functions? Will we basically be writing a custom, optimized set of graphics functions for use with VGA? We'll have to implement something like adaptive tile refresh to make it scroll smoothly.

Either hard-code graphics routines directly, or use mode specific drivers (EGA/VGA/whatnot) so that only one needs to be loaded at a time to save some memory. But there's the expense of calling subroutines through a pointer. Many games have e.g. CGA graphics set and driver, the main code that calls graphics does not care what the mode is.

keenmaster486 wrote:

How should graphics be drawn? We could just have the graphics as BMP files (the tiles in sheets, the sprites individually) and whoever does the graphics can use their tool of choice. I also developed a custom sprite editor once that I really like 😀 I could modify it to load/save in whatever format is required. But the tiles, I think, ought to be large BMP files pasted together from various sources. I assume we're using 16x16 tiles.

That would work for starters, and in real mode the easiest limit is to fit a tileset into 64kilobytes, which means 256 tiles of 16x16 size at 8-bit palette. You can of course have many tilesets, like per-level tileset or foreground/background tileset.

What about music? Sound effects? Back in the day Adlib/EGA or SoundBlaster/VGA would have been the de facto set.

There's plenty of good EGA pixel art by Surt to prototype with as well.

Also how are the real-mode tile limits for a 640x200x4 / 640x400x4 game?

Well 64 kilobytes gives you 512 tiles of 16*16 pixels at 4bpp.

First EGA cards had only 64 kilobytes of memory, thus capable of displaying only up to 640x350 resolution at 4 colors. So only 320x200 available at 16 colors, with 2 video pages. There was a 64 kbyte add-on card, but many later cards came with 128k.

I recently reaccustomed to the dither painting in PSP5. surely there must be a better modern solution but this works for now. This is more or less the art style with the video modes I had in mind (she's slightly squished for 640x400)

or the 640x200 PC88->EGA translation job kind

These characters are just practice/examples.

VileRancour wrote:

From what I know, OpenWatcom is the way to go for protected mode. For pure real-mode, maybe a 'classic' borland compiler with performance-sensitive bits converted to assmebler as you go.

During development of 8088 MPH I switched from Turbo C++ 3.1 to OpenWatcom 1.9, because I found it to be nicer to work with (more modern C/C++ support, nicer inline assembler, cross-compiling from Windows), and it also generated considerably better code than Turbo C++ did.
So I can recommend OpenWatcom for real mode/8086/8088 stuff.
As an extra bonus, it's open source, so you can study its libraries and tools and modify them if you see fit (I had to fix a bug with FPU detection on 8088).

leileilol wrote:

This is more or less the art style with the video modes I had in mind

Works well over my expectations! Always loved your character style, and I like your use of dithering and color palettes. Very vibrant.
Do you have examples of sprites for smaller characters too? (for example, top-view (ex.: Zelda : A Link to the Past) or side-view (ex.: Super Mario World), that could be used for gameplay segments?

Also, when we really get to development in the future, we might want to split foreground art (characters, objects) from background art (scenery) to two different artists. The goal being to not overwhelm a single person with everything at once, and incidentally have two slightly different art styles so the player becomes accustomed to what is / is not interactable in the world. Not necessary for now though 😀

This said, leileilol, out of curiosity do you have any examples of more still-life stuff that you can show?

(double post)