Double Dragon - Arcade port in C

[mcgeezer]

Those who know will understand that making Amiga games to a decent standard under the hardware limitations is a hard process that requires a lot of time, will power... an understanding wife too in my case.

Double Dragon is an iconic game and the beef I had with the developer was that the project was made to look as if it was easy to achieve with maybe a touch of a cavalier attitude.

But on saying that, Amiga game developers owe nothing, we're just trying to have fun with the machine we love.

I'm 100% sure Double Dragon can be made arcade accurate (not perfect) on an A500 with a 1mb expansion, someone doesn't have to spend a year of their time proving it though. You've got Metro Siege coming... that game will be a big hit.

[jotd]

^^ that

Plus when you're using an emulator to code your game, don't wait too long before testing on real hardware because chipmem speed can make a great difference. Specially when using an engine or API (Scorpion, SDL...)

I had Xevious flying on vanilla A1200 speed, and crawled/flickered on real A1200/060/fastmem. Big disappointment (but a big issue was fixed afterwards fortunately)

[Tigerskunk]

Retro Game Dev is hard.

And, yes, devs don't owe to anybody what they do in their spare time.

But: It's also tiresome in my opinion if guys post their "Hey, I worked a day on a Scorpion Engine version of FAMOUS ARCADE GAME and got those assets from spritedatabase.com to run around and jump" here and on FB and other places to generate hype once a week, yet never to be seen again afterwards after they got their mention on Indie Retro News.

Poisons the waters for those few people who are a bit more serious on Amiga game dev and who tend to finish things because it gives off the idea that Amiga game dev is piss easy now, and there are a hundred games on the way in this moment.

Yet the truth is, it's still only a few people finishing creating games.

YMMV off course, this is just my 5 ct.

[DanielAllsopp]

Originally Posted by jotd

^^ that

Plus when you're using an emulator to code your game, don't wait too long before testing on real hardware because chipmem speed can make a great difference. Specially when using an engine or API (Scorpion, SDL...)

I had Xevious flying on vanilla A1200 speed, and crawled/flickered on real A1200/060/fastmem. Big disappointment (but a big issue was fixed afterwards fortunately)

Exactly this, I've been caught out a few times with this which necessitated lots of design and code changes. Now, every time I finish a milestone I run a little script which writes and ADF and dumps it onto a floppy disk with my GreaseWeazel:



They're then chucked straight into an A4000/060 and a bog standard A1200 and tested side-by-side with the emulator. More often than not everything's fine, but there are some gotcha moments to look out for!

As you can see, I've not looked at my code for about 6 weeks now. Doesn't mean it's cancelled, it just means I haven't had the physical time to do anything lately but I will pick it up again soon. Believe me, it's way more interesting writing 68k code than writing Swift code so it's not through a lack of interest. The hardware challenges are great too, well most of the time

[jotd]

Originally Posted by Tigerskunk

Retro Game Dev is hard.

It's also tiresome in my opinion if guys post their "Hey, I worked a day on a Scorpion Engine version of FAMOUS ARCADE GAME and got those assets from spritedatabase.com to run around and jump" here and on FB and other places to generate hype once a week, yet never to be seen again afterwards after they got their mention on Indie Retro News.

Yet the truth is, it's still only a few people finishing creating games.

YMMV off course, this is just my 5 ct.

I couldn't agree more! Daniel don't feel targeted because you're not

When I was younger I was starting a lot of things I didn't finish. Now I tend NOT to start things if I feel I can't finish them. That works most of the time (you are entitled to a few jokers). The trick for me is to have a plan about the game features. Not start something without having thought how to encode levels, etc... For instance I wouldn't take on a tiled platformer game without the maps & level layout (items) available (and also, because I don't want to code a 8 way scroller with tiles on the amiga, it looks difficult )

[acidbottle]

Talking from a Scorpion development point of view. It is SO EASY, to get up and running, the encouragement brings motivation, which, especially with arcade ports is great for a while, then you hit walls that require quite a bit of out the box thinking, then developers give up. With asm and the likes there is always a solution, which will likely require some skills to get around. The solutions in Scorpion are often difficult, or not possible because it is physically incapable of doing everything, there is a definite finite nature to what you can realistically achieve with it, though this goal post moves frequently, same with Redpill I imagine.

Have started a few projects with Scorpion, I really want to disclose one particular project's progress but is no point as all the above comments are correct, get the game running to a point it is 100% possible to complete, with time and energy, then wet the appetite, so to speak. It's close by though

Is Double Dragon achievable in this Engine, eventually I would say yes. As McGeezer says though, Metro Siege is on the horizon and should satisfy the brawler department for a while. Can't wait for that!

[DanielAllsopp]

Originally Posted by jotd

I couldn't agree more! Daniel don't feel targeted because you're not

Nah, I don’t feel targeted mate, far from it. I just wanted to point out my particular case. It’s different for all developers I guess. All is well here.

[gimbal]

Originally Posted by mcgeezer

Double Dragon is an iconic game and the beef I had with the developer was that the project was made to look as if it was easy to achieve with maybe a touch of a cavalier attitude.

It is usually when I reach the halfway point in a game project that I realise "yep, this is about ten times more work/difficult than I imagined it to be" It is usually very easy to get a game project going if you know at least something about the platform/tooling. You borrow some stuff from other projects and hammer that into the framework that you need for your new game. And how fast that goes can easily mislead you into thinking the rest of the development process is going to be equally speedy.

As an example if you're creating a horizontal or vertical shooter, the first think you will likely want to achieve is have a ship flying with a scrolling background. Might not take too long, you might use placeholder graphics. And then it is easy to think "Hah! Halfway done!". Nuh-uh. 5% done.

[malko]

Originally Posted by jotd

[...] (and also, because I don't want to code a 8 way scroller with tiles on the amiga, it looks difficult )

Fortunately some nice guys leave some source code available behind them and it can become a great source of knowledge for things to be less difficult
https://night.owl.de/index.php?id=23

[jotd]

True, but you often find yourself removing all the code you don't want and understanding the code you want to keep and it's sometimes pretty daunting.

I built all my blitter & sprite routines from bits copied & modified from a lot of sources, but now they're mine, and I reuse them for all my games. The problem with those routines is that you have to understand the interface with them. It's not like an unpacking routine, which is opaque but has a simple interface.

I think that if I did a 8 way scrolling game, unless I'm going Scorpion, I'd consult the sources on aminet (which are in C) and would try to roll my own, seeking help from ross, a/b, roondar, mcgeezer, and other helpful people out there.

In the case of Double Dragon, this is not really 8-way. The horizontal dimension is 'infinite' (means you cannot allocate all the memory at once for all the playfield), but the vertical dimension is pretty limited and not that high (maybe twice a screen?). It means that you can handle that like a horizontal scroller. Maybe it's going to be underperformant & wasteful because you'll have to draw all the tiles on the whole height by doing so (even if you don't visit all the squares), but at least it's simple. And for unreachable areas, you can have a "blank" tile that is never drawn so it's zero cost.

On the other hand, Shinobi starts pretty much horizontal, but you realize it's 8-way in the later levels. So if you designed your scrolling to be finite vertically, you'll have to redo everything from scratch. That's one of the elements of the planification that you must take into account when starting a project. What's easy now is going to be unuseable later, unless you give up your project in the middle

[Tigerskunk]

Originally Posted by DanielAllsopp

Nah, I don’t feel targeted mate, far from it

And indeed you shouldn't, since you went beyond that point by a good margin already, imo.

[Tigerskunk]

Originally Posted by acidbottle

Talking from a Scorpion development point of view. It is SO EASY, to get up and running, the encouragement brings motivation, which, especially with arcade ports is great for a while, then you hit walls that require quite a bit of out the box thinking, then developers give up. With asm and the likes there is always a solution, which will likely require some skills to get around. The solutions in Scorpion are often difficult, or not possible because it is physically incapable of doing everything, there is a definite finite nature to what you can realistically achieve with it, though this goal post moves frequently, same with Redpill I imagine.

I guess tailoring the game around the engine's capabilities would be easier then...

Originally Posted by acidbottle

Have started a few projects with Scorpion, I really want to disclose one particular project's progress but is no point as all the above comments are correct, get the game running to a point it is 100% possible to complete, with time and energy, then wet the appetite, so to speak. It's close by though

Looking forward to what you are up to, mate...

[mcgeezer]

Originally Posted by jotd

On the other hand, Shinobi starts pretty much horizontal, but you realize it's 8-way in the later levels. So if you designed your scrolling to be finite vertically, you'll have to redo everything from scratch. That's one of the elements of the planification that you must take into account when starting a project. What's easy now is going to be unuseable later, unless you give up your project in the middle

This exact scenario happened when i was doing Rygar. The game is a bit deceptive in that it looks like it only scrolls horizontally... until you get to level 5 when the vertical scrolling is required when you have to climb the rope. It was by shear luck that during that scene hardly anything goes on and I was able to use the pristine buffer to update the front and back buffer fully using the blitter every frame without any slow down... very very lucky, but even then it was damn tricky because I neglected to think about it up front.

Incidentally after I coded Rygar I did make a working efficient 8 way scroller in assembler which I've not put to any use yet.

[gimbal]

Double Dragon 2 demonstrates that as well to some extent, just not in the Amiga version. If you play the arcade version then the transition to the next level is some kind of elevator which causes vertical scrolling. They cut that out in the Amiga version to be able to keep it a pure horizontal scroller I guess.

[Havie]

Originally Posted by mcgeezer

This exact scenario happened when i was doing Rygar. The game is a bit deceptive in that it looks like it only scrolls horizontally... until you get to level 5 when the vertical scrolling is required when you have to climb the rope. It was by shear luck that during that scene hardly anything goes on and I was able to use the pristine buffer to update the front and back buffer fully using the blitter every frame without any slow down... very very lucky, but even then it was damn tricky because I neglected to think about it up front.

Incidentally after I coded Rygar I did make a working efficient 8 way scroller in assembler which I've not put to any use yet.

Rastan...

[mcgeezer]

Originally Posted by Havie

Rastan...

Isn't someone already doing it? I'm sure I seen that on twitter not so long ago.

[jotd]

acidbottle started something with Scorpion

https://eab.abime.net/showthread.php?t=110469&page=2

Originally Posted by mcgeezer

This exact scenario happened when i was doing Rygar. The game is a bit deceptive in that it looks like it only scrolls horizontally... until you get to level 5 when the vertical scrolling is required when you have to climb the rope. It was by shear luck that during that scene hardly anything goes on and I was able to use the pristine buffer to update the front and back buffer fully using the blitter every frame without any slow down... very very lucky, but even then it was damn tricky because I neglected to think about it up front.

So it happened to someone after all! I can't imagine refactoring a game to change the scrolling to 8 way in the middle of the development.

[Konrad]

Originally Posted by gimbal

Double Dragon 2 demonstrates that as well to some extent, just not in the Amiga version. If you play the arcade version then the transition to the next level is some kind of elevator which causes vertical scrolling. They cut that out in the Amiga version to be able to keep it a pure horizontal scroller I guess.

Double Dragon 2 does scroll vertically, though:
see 7:20 / 7:48 and especially 13:30
[ Show youtube player ]

[gimbal]

I don't see any at 7:20 (besides the little bit of vertical scrolling that is throughout the entire game), but indeed at 13:30 there is a ladder going down. Huh.

Good god, what music during gameplay. I'm not sorry I never figured out back in the day that you can apparently toggle it.

[acidbottle]

Originally Posted by Tigerskunk

I guess tailoring the game around the engine's capabilities would be easier then...


Looking forward to what you are up to, mate...

For sure, once you get familiar with the engine you soon become aware of some limitations, usually after prototyping! That said there are some clever people using it out there and some proper programmer knowledge, as I discovered, also helps tremendously!

Cheers dude, have a new found respect for anyone making shooters though, it is some of these mechanics that are holding me back a bit. They can certainly be a bit tricky to fathom out

@jotd, that was a classic case of prototype easy, full game .. lots of work haha! Not from me, though I believe Rastan fans though may eventually one day get a nice surprise ...