Programming the 7800

[philipj]

I think it's a great idea to have some kind of definitive website

for the Atari 7800 programmer. I don't have the skills yet, but I

like to make games for the 7800 system. I had to look almost

everywhere to get most of the programming info until I ran into

the Atari Historical Society. The potential is there for new

developments let's strike the iron while it's hot.

[caveman]

I'd like it for someone could come up with a hack for Jinks, at least one

to make the sounds less grating, or maybe make the psychedelic backgrounds easier to look at, or maybe create a new one altogether. I always thought that it was a cool concept of a game, just that the end product was so shitty.

[DracIsBack]

I don't mind JINKS as much as others. I mean, any game that manages voice synthesis on the 7800 without a POKEY can't be all bad!

[philipj]

Has anyone been able to use any other add-on sound chip other than a Pokey? I beleive that a good way to start a new thing with the Atari 7800 development is to try something new. What do you all think?

[EricBall]

Sure, stick a SID (6582) in a 7800 cart. I don't see any reason you couldn't.

[Vigo]

Several reasosns why it wont work: it needs 2 power currents (5V and 9/12V (depending on the model)) and its maximum clock frequency is 1Mhz. (Btw the model you mentioned is the most hardest to find... )

 

 

Better would be something like an OPL2, OPL3 or OPL4 chipset, but i guess it would be too much CPU overhead generating sounds with these devices...

[EricBall]

Several reasosns why it wont work: it needs 2 power currents (5V and 9/12V (depending on the model)) and its maximum clock frequency is 1Mhz. (Btw the model you mentioned is the most hardest to find...  )

 

Whoops, missed the Vdd input and the CLK max. In that case you are correct.

 

But, any memory-mapped sound generator with TTL level analog output should be okay.

[EricBall]

Just to wrap this topic up.

 

The general consensus on the Stella mailing list was it was okay (with most people) to use it also for 7800 programming topics, as long as this is indicated in the subject. (I recommend starting the subject with MARIA since the biglist search engine doesn't handle numerics.) The assumption is the volume will be low enough not to be a big problem. Plus it might encourage some of the existing 2600 homebrewers to give the 7800 a try.

 

On a personal note, I'm going to put my 7800 programming on hold until I get my Lode Runner style game done. I just don't have enough time to do both and I don't want to let Leprechaun sit partially completed while I play with the 7800.

[+Propane13]

The general consensus on the Stella mailing list was it was okay (with most people) to use it also for 7800 programming topics, as long as this is indicated in the subject.  (I recommend starting the subject with MARIA since the biglist search engine doesn't handle numerics.)  The assumption is the volume will be low enough not to be a big problem.  Plus it might encourage some of the existing 2600 homebrewers to give the 7800 a try.

 

Hmm.. that'll get you the perspective of 2600-programmers, yes, but what about the 7800 community? Maybe they do not wish to infiltrate the stella list. Personally, I do not wish to re-join stella list. I left mainly because of 50+ emails a day, and most that were not relevant. Bin files fly in a lot, and every week there's the "hi, I'm a newbie" thread. Most conversation is just rehashing of things that could have been found by simply browsing the archives. I grew tired of filtering most of the emails into the trash.

 

Note that a 7800 mailing list is something that I definitely do NOT want to miss out on. If I need to rejoin stellalist, I'll consider it, but I strongly object. Of course, this may be simply limited to my unique perspective and experience. Does anyone else feel this way?

 

How much does an account at biglist count? I *may* be interested in helping to fund a separate 7800-biglist account, depending on price, and usage of the community-- but, if only 5 people use it, it may be easier (and cheaper) to grow this list as a simple email list without any real biglist archives until more people are on board. All that is really needed is a moderator. Archives can be kept later in the 7800 lifecycle, because I don't think they're too necessary at this stage. Would anyone be interested in doing it this way? That way, it can start small, and grow when needed. I have a fear of this dying out on the stella list, or of people changing the subject of these emails to make them 2600-related. It's too big, and the 7800ers are too few. I don't think it makes sense.

 

Opinions?

 

Thanks,

-John K. Harvey

[vanman]

If you get me a tutorial then I'm in.

[Vanguard]

Heck, if there was some way to get started, I would love to get some classic game development going for the 7800. There's a lot of classic games from that era and previous that I would like 'brought up' to the 7800's capabilities.

 

Trouble is, there isn't a lot to get started. I'm also a long-time C/C++ programmer that has absolutely no desire at all to go back to assembly programming. Are there even libraries for the 7800 yet?

[EricBall]

It might be possible to use C to generate code for the 7800. I'm finding coding for the 7800 to be mostly data movement work. And other than the DLL NMI, there isn't anything timing sensitive (though it's real easy to chew through the VBLANK time: building the DLL takes time).

 

But ASM ain't that tough to learn if you already have a decent handle on using pointers in C. And once you learn ASM, you can start to see how a C compiler generates code.

 

I don't know about whether it would be feasible to create a C library for the 7800. Maybe some headers. You'd probably still have to have the C compiler generate ASM files and tweak those as necessary. Maybe a 7800 specific C compiler.

[Vanguard]

But ASM ain't that tough to learn if you already have a decent handle on using pointers in C.  And once you learn ASM, you can start to see how a C compiler generates code.

 

I know how to do assembly, I just don't desire to return to it. I would imagine that a C library would be a C function interface to assembly calls, which wouldn't bother me at all. But using assembly to code the logic of the gameplay just... gives me unpleasant flashbacks.

[DEBRO]

This may have been covered before but...

 

Does anyone one know why Atari went with 4K of RAM for the 7800 when the 5200 had 16K? It seems RAM would have been cheaper by the time the 7800 came out. I know you can add RAM to indivdual carts but I just wonder why they went this route

[DracIsBack]

Weirdly enough, Nintendo did the same with the NES - 2K RAM + 2K video RAM.

 

Perhaps they thought it made more sense to add the memory to cartridges as specifically needed?

[DanBoris]

This may have been covered before but...

 

Does anyone one know why Atari went with 4K of RAM for the 7800 when the 5200 had 16K? It seems RAM would have been cheaper by the time the 7800 came out. I know you can add RAM to indivdual carts but I just wonder why they went this route  

 

It may have been a cost issue. The 5200 used dynamic RAM which took up a lot more space on the board then the static RAM that the 7800 used. This obviously wasn't an issue with the 5200, they had board space to spare, but the board in the 7800 is packed pretty tight as it is.

 

They also may have felt that they didn't need it. On the 5200 there needs to be a block of memory set aside for the screen buffer. In the higher resolution modes this buffer required almost 8K for a single non-scrolling screen. The 7800 on the other hand doesn't need a screen buffer and it's graphics data strucutres take up far less memory, so you can generate complex displays with a lot less RAM.

 

Dan