Theoretical approach to Tempest on the 2600

[Warboss Gegguz]

Been delving into my collection of protos and homebrews, Specifically the partially fixed version of Tempest by CDS games. It's better, but it always makes me try to think of a way in which Tempest on the 2600 could've been converted in a less compromised way.

The way I see it, the biggest issue is just the amount of sprites and real estate allocated to the levels themselves. However, I've considered the idea of substituting the solid rails of tempest's typical levels with a dot matrix akin to something like Beamrider, but topped with solid rails to properly convey the navigable portions of the levels.

Note, I'm not necessarily looking for it to look perfectly like tempest. Ideally it should be as close as possible, but the gameplay takes priority.

 

This is more just a thought experiment, as while I love game design I have basically 0 programming skills... Though if anyone's interested, feel free. I just feel like there has to have been SOME WAY to do it without butchering it beyond playability. Especially when there are plenty of similar games even from the same era that work on the console.

[Lord Thag]

I agree with you, there is certainly a better way.

 

Look at the homebrew Star Castle. That was supposed to be impossible too (in fact, the original attempt became Yar's Revenge). Or Boulderdash. Or the full Donkey Kong homebrew port. I'm sure it's doable these days but doable and likely to be done are two different things. It takes a very talented coder to pull off something that completely breaks what the old 2600 is supposed to be able to do well enough to be playable.

 

If we ever do get a better port, I suspect it will be champ games and use that additional ARM chip.

[Warboss Gegguz]

On 12/5/2023 at 9:49 AM, Lord Thag said:

I agree with you, there is certainly a better way.

 

Look at the homebrew Star Castle. That was supposed to be impossible too (in fact, the original attempt became Yar's Revenge). Or Boulderdash. Or the full Donkey Kong homebrew port. I'm sure it's doable these days but doable and likely to be done are two different things. It takes a very talented coder to pull off something that completely breaks what the old 2600 is supposed to be able to do well enough to be playable.

 

If we ever do get a better port, I suspect it will be champ games and use that additional ARM chip.

Again, I don't really have much programing skill, but I wholly understand why converting a vector game to a raster display let alone on 80s tech, is extremely difficult.

The only real boon to tempest is that it obviously is a mostly black screen, but again I feel when you're going from Arcade to far less powerful hardware the main goal is a proper facsimile of the gameplay rather than a facsimile of the visuals. The most obvious examples to me of a complicated vector game with a solid port would be Battlezone.

Does it look like its arcade counterparts Somewhat. Is the gameplay accurate to the arcade. Absolutely. And some of the compromises to visuals (like the 3rd-ish person perspective) are done for the sake of gameplay. The only real shortcoming is the lack of obstacles.

 

And I think games like Tempest could be done in a similar way, especially with how far optimization has come... But again, I'm a designer/playtester, not a programmer. So I can only throw out ideas with my understanding of the tech.

 

On that note, the control scheme in the prototype is perfect in terms of human factors and console limitations, and would honestly be decent were it not for the jankiness around hitboxes and bullet paths even in the "fixed" versions. I just think that if you're going to start redoing this, it would be better to start from square 1. It's so basic that it only really serves as a proof of concept.

 

I'd imagine it would only be able to be the first stage or maybe 2 stages if we're REALLY pushing it and just loop with increasing difficulty/enemy variety. I just feel like I've seen equally complex games pulled off on the 2600 even back in the 80s, albeit not from atari.

[Lord Thag]

22 hours ago, Warboss Gegguz said:

Again, I don't really have much programing skill, but I wholly understand why converting a vector game to a raster display let alone on 80s tech, is extremely difficult.

The only real boon to tempest is that it obviously is a mostly black screen, but again I feel when you're going from Arcade to far less powerful hardware the main goal is a proper facsimile of the gameplay rather than a facsimile of the visuals. The most obvious examples to me of a complicated vector game with a solid port would be Battlezone.

Does it look like its arcade counterparts Somewhat. Is the gameplay accurate to the arcade. Absolutely. And some of the compromises to visuals (like the 3rd-ish person perspective) are done for the sake of gameplay. The only real shortcoming is the lack of obstacles.

 

And I think games like Tempest could be done in a similar way, especially with how far optimization has come... But again, I'm a designer/playtester, not a programmer. So I can only throw out ideas with my understanding of the tech.

 

On that note, the control scheme in the prototype is perfect in terms of human factors and console limitations, and would honestly be decent were it not for the jankiness around hitboxes and bullet paths even in the "fixed" versions. I just think that if you're going to start redoing this, it would be better to start from square 1. It's so basic that it only really serves as a proof of concept.

 

I'd imagine it would only be able to be the first stage or maybe 2 stages if we're REALLY pushing it and just loop with increasing difficulty/enemy variety. I just feel like I've seen equally complex games pulled off on the 2600 even back in the 80s, albeit not from atari.

Look at 2600 Star Castle, it's a vector game. The technique used for the central forcefields could probably be used to do the level graphics for tempest.

 

I am a (hobbyist) programmer. I'm absolutely not an expert on the 2600, but I did program a working demo on it once, so I know the basics of how the system works. It's like this: The VCS has two player sprites (which can be doubled and tripled), two missile sprites, one ball (bigger missile) and a playfield (the latter having  three pixels wide elements, which is why 2600 games have lines/wider sprites on the level graphics), as well as built in collision detection. The system only has a *scan line* (literally one line of a TV display) of video memory, where you literally have to reconfigure your logic between the horizontal blank time (when the old school TV switches off its beam when it's reached the right of the screen and moves it back to the left) and then resumes. So lets say you draw a ship at location X on one line, and you also want a different graphic on location Y on the same line using that sprite, if your program logic is fast enough to redo the graphics on the sprite and update it's location, the Atari is too 'dumb' to know it's already drawn it and will happily draw it again. Or you can draw one sprite in one frame, and the second in another (the reason for so many games flickering).

 

The Atari actually has great, built in hit detection between players and missiles, and players and the playfield. So it wouldn't be janky if programmed correctly. I don't think the old Tempest proto is a great example of what could be done, frankly.

 

I suspect the solution would be to use player or missile graphics and very fast/clever usage of drawing different level 'lines' every other frame, the built in sprite ability to double or triple up, and so on. That would require a very tight kernel to do all that and have enemy sprites continuously spawn. There are plenty of folks here with that level of talent though. I suspect it would be much less problematic with the additional ARM chip Champ games uses too. Their games frequently break what was previously thought to be possible on the VCS with that ARM chip handling things. I think Tempest would be *very* doable with that. And with the partnership between Atari/Atariage now being a thing, rights would not be a problem now either.

 

Fingers crossed we'll see it some day. If they can do Galaga, Elevator Action, Bosconian and Star Castle now, they can do Tempest.

[BIGHMW]

ABSOLUTELY!!! If they can do Tempest the justice it deserves than I’d say that the 2600 would be the one if it came down to one console I’d keep.

 

40 years ago I would’ve said the 5200 but the graphics and how homebrewers have been able to squeeze out each and every one of those 128 bytes of RAM nowadays in a 2600 is quite impressive!!!

[BillDMatt]

Champ Games I think could do it,   make it work with the Indy 500 paddle   that would be awesome. 

[davyK]

This has been discussed before...