Atari VCS and Apple II developer environment? What did they do?

[Keatah]

You can always read about how the Apple II was used in development of VCS games, but that's about as much as you learn.

 

So how did Atari use it? Did they simulate the TIA on it? Was there any direct electrical connection between the Apple II and the VCS? Maybe they just used the Apple II as a terminal or storage device.. Or maybe to test and trace 6502/6507 code?

 

Maybe they just used it as a utility computer, to convert hex/dec and build up playfields and charts and sprites and stuff like that.

 

Though I can't imagine the Apple II producing anything as colorful or "soundful" as a VCS game - the Apple II somehow played an important part in writing VCS software back in the day. What was that role?

[+Andrew Davie]

You can always read about how the Apple II was used in development of VCS games, but that's about as much as you learn.

 

So how did Atari use it? Did they simulate the TIA on it? Was there any direct electrical connection between the Apple II and the VCS? Maybe they just used the Apple II as a terminal or storage device.. Or maybe to test and trace 6502/6507 code?

 

Maybe they just used it as a utility computer, to convert hex/dec and build up playfields and charts and sprites and stuff like that.

 

Though I can't imagine the Apple II producing anything as colorful or "soundful" as a VCS game - the Apple II somehow played an important part in writing VCS software back in the day. What was that role?

 

 

Interesting questions.

Firstly, the Apple has a keyboard, so the code would definitely have been written on this machine. It would have also been assembled -- pretty straightforward as it's also a 6502-based platform, so you could probably use the native assemblers for Apple II to assemble 2600 code. So, edit/assemble cycle on the Apple II -- and I bet my bottom dollar that they had a cart emulator that plugged into the 2600 and connected to the Apple -- and you'd download the binary to it, hit reset on the '2600 and test your program on actual hardware. We used a similar setup for C64 programming and NES programming back in 1986 or so; using the BBC as the development machine, and we'd download the binary to the cart emulator and test on actual hardware. Fairly efficient development cycle.

[Retro Rogue]

You can always read about how the Apple II was used in development of VCS games, but that's about as much as you learn.

 

So how did Atari use it? Did they simulate the TIA on it? Was there any direct electrical connection between the Apple II and the VCS? Maybe they just used the Apple II as a terminal or storage device.. Or maybe to test and trace 6502/6507 code?

 

They didn't. 2600 development at Atari was usually done on 6502 simulators running on mainframes and then code was burned to eproms and tested on an actual 2600. Here's a picture of Steve Woita during the development of Taz/Asterix in 1983.

 

 

 

In relation to Andrew's comment regarding the Apple having a keyboard right on it which would make it easier for development, that's exactly one of the form factors Jay Miner and Joe Decuir wanted for their next project at Atari in '79 before they quit. Specifically they wanted to do a 68000 based console/computer that had a keyboard on it already so programmers could code games right on the console and not have to do the whole eprom burn thing to test. They carried that idea on to Amiga when their Lorraine was initially a game console with a keyboard attached (think the later Amiga 500).

Edited October 3, 2013 by Retro Rogue

[+SvOlli]

I also did some research on "how did they develop back then" some time ago, so I want to throw in some words of mine as well.

 

They didn't. 2600 development at Atari was usually done on 6502 simulators running on mainframes and then code was burned to eproms and tested on an actual 2600. Here's a picture of Steve Woita during the development of Taz/Asterix in 1983.

 

This is not quite right. The eproms were not used for development, more for demonstrating prototypes. I once asked Warren Robinett about the tools of development, since this is missing in his slides about the creation of Adventure. Here's the part from his reply:

 

 

There was, as you guessed, a board with RAM where the ROM normally sat in the address space of the system, into which the assembled machine code code be downloaded. It actually had 8K bytes of RAM, so you could run a program bigger than would fit in the (non-bank-switched) 4K maximum address space of the 2600. Trying to reduce the program size to fit into the 4K bytes available in the ROM was called crunching.

[Retro Rogue]

I also did some research on "how did they develop back then" some time ago, so I want to throw in some words of mine as well.

 

 

 

This is not quite right. The eproms were not used for development, more for demonstrating prototypes. I once asked Warren Robinett about the tools of development, since this is missing in his slides about the creation of Adventure. Here's the part from his reply:

 

Actually that's not quite right either. It depends on the generation and what was readily available. We talked to many 2600 programmers from the '76 to '84 period, and a lot of them talked about the exact process I just mentioned, including the frustration of needing to keep running to the burning room to wait and burn an EPROM and run back and test it. Nothing to do with demo of prototypes at that time. Later they got the RAM boards you're talking about.

[+Andrew Davie]

Actually that's not quite right either. It depends on the generation and what was readily available. We talked to many 2600 programmers from the '76 to '84 period, and a lot of them talked about the exact process I just mentioned, including the frustration of needing to keep running to the burning room to wait and burn an EPROM and run back and test it. Nothing to do with demo of prototypes at that time. Later they got the RAM boards you're talking about.

 

This is true for me, too. In the early NES days we didn't have a downloadable system. I had to burn EEPROMS to test the code changes. It was about a 15 minute wait, IIRC.

We had a rotation going; EEPROMS erasing in the UV enclosure, EEPROMS in the cart we were testing, and EEPROMS ready to burn. It was slow and frustrating!

[+Gemintronic]

Atari never expected third parties but Nintendo made gobs of money off of them. They didn't have tools that came with the pricey license?

[+Andrew Davie]

Atari never expected third parties but Nintendo made gobs of money off of them. They didn't have tools that came with the pricey license?

 

We reverse-engineered the NES, so didn't have the manuals or tools that Nintendo provided to licensed developers.

[Wickeycolumbus]

There's the READS development system from VSS that hooks up to the joystick ports of an A8 computer and downloads a game to a RAM cartridge:

 

http://www.digitpress.com/library/interviews/interview_ed_salvo_9.png

 

(Taken from this interview)

 

I remember seeing something posted about an Apple II development system, I believe it was called the frob or something like that. Can't seem to find anything about it though. Using home computers was probably common among the smaller developers and 3rd parties in general for the 2600. For example, the guy that programmed Red Sea Crossing used a Franklin Ace and an EPROM programmer. Making a modern system using an Apple II or similar computer would be a pretty cool/interesting project.

 

Edit: found the frob post:

 

http://atariage.com/forums/topic/134938-frob-and-frob-52-pictures/

Edited October 4, 2013 by Wickeycolumbus

[Nukey Shay]

BTW Arcadia/Starpath games were cross-developed using AppleII computers. Their Supercharger hardware was designed to be directly compatible to the computer's native cassette port signal.

[Retro Rogue]

There's the READS development system from VSS that hooks up to the joystick ports of an A8 computer and downloads a game to a RAM cartridge:

 

http://www.digitpress.com/library/interviews/interview_ed_salvo_9.png

 

(Taken from this interview)

 

I remember seeing something posted about an Apple II development system, I believe it was called the frob or something like that. Can't seem to find anything about it though. Using home computers was probably common among the smaller developers and 3rd parties in general for the 2600. For example, the guy that programmed Red Sea Crossing used a Franklin Ace and an EPROM programmer. Making a modern system using an Apple II or similar computer would be a pretty cool/interesting project.

 

Edit: found the frob post:

 

http://atariage.com/forums/topic/134938-frob-and-frob-52-pictures/

 

 

Yes, there was a number of third party 2600/5200 dev systems for both the 800 PCS and Apple II at the time. They were frequently advertised at the back of computer magazines.

[Mr SQL]

 

We reverse-engineered the NES, so didn't have the manuals or tools that Nintendo provided to licensed developers.

Andrew,

how did your company get around Nintendo's safeguards? (lockout chip, getting sued, etc)

Were there attempts to obtain or create a RAM board to speed the development process?

[+Andrew Davie]

Andrew,

how did your company get around Nintendo's safeguards? (lockout chip, getting sued, etc)

Were there attempts to obtain or create a RAM board to speed the development process?

 

We weren't a publisher, so we developed games for publishers who WERE licensed. That's how we avoided being sued. In the end, our knowledge of the NES was probably better than Nintendo's, and we were doing stuff with the machine they couldn't. At some point, we just became an accepted developer and were sent Nintendo's documentation on the machine which, I recall, was nowhere near as comprehensive as our in-house stuff. For example, we could do split-screen 8-direction scrolling on the basic machine.

Memory fails me on the lockout issue. We developed for very basic cart formats with no lockout? MMC1....? It's been a long long time. I know I never had to worry about lockout chips.

Yes, we had an in-house developed RAM board for speedy development.

Cheers

A

 

[Mr SQL]

 

We weren't a publisher, so we developed games for publishers who WERE licensed. That's how we avoided being sued. In the end, our knowledge of the NES was probably better than Nintendo's, and we were doing stuff with the machine they couldn't. At some point, we just became an accepted developer and were sent Nintendo's documentation on the machine which, I recall, was nowhere near as comprehensive as our in-house stuff. For example, we could do split-screen 8-direction scrolling on the basic machine.

Memory fails me on the lockout issue. We developed for very basic cart formats with no lockout? MMC1....? It's been a long long time. I know I never had to worry about lockout chips.

Yes, we had an in-house developed RAM board for speedy development.

Cheers

A

Very cool your team was able to push the hardware, learning beyond Nintendo spec - split screen 8-way scroll is pretty awesome! Seems the Nintendo graphics chip had undocumented/unexpected functionality waiting to be explored, just like the TIA.

 

Regarding RAM boards for the VCS, I think by 1982 all 3rd party developers who didn't have them already could have used the SuperCharger as a basic 4K board without knowing about it's unique features (RAM access method, etc).

[tschak909]

I remember a Frog Development house that advertised an Apple 2 dev system for VCS dev, that was $495, and included a ram cart, an interface card, and basically a pamphlet for a manual.... Advertised in the back of creative computing or Interface Age circa 1983...

[enthusi]

When did the dev-kits come up btw?

I think there are some around for the 7800 at least?

[Retro Rogue]

When did the dev-kits come up btw?

I think there are some around for the 7800 at least?

 

Different Atari company (Atari Corporation). Most of Atari Corp.'s game development was by outside parties, hence the need for dev kits.

[DanOliver]

That picture of Steve Woita has a big box to the left of the TV that I think is what the VAX downloaded code into. I don't think I ever got to use one. I did write 6502 on Atari's VAX I think for the 800...but can't remember.

 

At Apollo we wrote code on an Atari 800 and used Atari's Assembler cart to compile. Then we burned EPROMs to test code. As I was leaving Apollo they had a hardware guy designing a RAM card type deal.

 

At VentureVision I wrote code on an Apple II using an assembler I wrote. I called it an instant assembler because the compile time was about 1 second. But I still had to burn EPROMs to test.

 

At Atari for Telepathy I used my Apple II system, with a 512K RAM drive (like $1000) using some RAM card in the 2600. So instant compile, instant download = sweet development system...just hope the power doesn't go off. The "some RAM card" I'll bet was a frob..thanks Wickeycolumbus, I'd forgotten about that. I remember it now as being the most unbelievably advanced piece of hardware ever created. Just seeing the name again made my knees instantly buckle. I doubt seriously you could have a better 2600 development system today using modern computers.

 

At Atari we had logic analyzers to help debug code. They weren't generally super helpful, but for certain bugs they were great. I think I was told they cost $60K.

[ianoid]

Chiming in:

 

I heard from Dan Kramer that they used a Cromemco system for some Atari development, but he didn't recall the specifics. He just said he recalled a large black box. I know Cromemco did S-100 computers, but I don't know about their mainframe legacy.

 

The Commavid company apparently wrote their 2600 games on a Coco (TRS-80 Color Computer). The genius who started the company reverse engineered the 2600 to do so. Prototypes that Kelly, Hardie and Santulli found there (well, at least one that was unreleased) were saved on cassette tapes- I think that's where Rush Hour came from.

 

The Frob- who wouldn't want one? A 2600 dev system for the Apple II. The one I know of is in Slapdash's hands. I had him write me into his will to leave it to me. MWAHAHA!

[DanOliver]

I've never really been interested in having old systems, not enough space. But I must say remembering frob...I now want one. And the Apple II was a beautiful machine to program.

 

The name Cromemco doesn't ring any bells, but I don't think I ever used one so I wouldn't know. I actually thought they were called Stella because I didn't even know the 2600 was called Stella until recently. Maybe someone called it Stella because there was a Stella in side. Would be interesting to find out what that box did. It was as big as a TV but seem to just act as a RAM card.

 

I learned the 2600 by reverse engineering and the manual from Commavid, so never saw the term "Stella".

[tschak909]

According to Joe DeCuir, Collen and Candy hardware was brought up on the Cromemco Z-2, which basically had most of the target hardware (6502, GTIA, ANTIC, POKEY) on an S-100 card.

[tschak909]

And yes, sorry, it was The Frob, that I was trying to remember...it has been...way.....wayy....waaaaaaaaaaaaay too long.

 

Wasn't it like, $495.00 ??

 

-Thom

[DanOliver]

Yes, $495 I saw in an ad someone posted. In those days $495 for something like that was a steal. And it was a very nice design and well built. At the time I think they might have been better off charging more, like $2-3K. It was such a limited market. Anyone writing a 2600 game then could sell the game for a lot of money so the cost of something like frob would be well worth a lot more.

 

Seems like maybe they got into the same issue as Magicard. Amazing value but I guess they thought tons of people wanted to program. At that time I know companies were willing to buy stolen 2600 programming manuals for hundreds of thousands of dollars. But then Magicard comes out for $50 with what turned out to be a complete 2600 programming manual. I think every company that started making 2600 games at the time bought one copy of Magicard. They could have priced it a little higher. Brilliant engineers to be sure, but could have used some sales and marketing.

 

Sure wish I'd kept my Apple IIe and frob card, I'd be writing a 2600 game right now. But sadly I tossed them. Just too much stuff. I'll have to take another look around, but I'm not hopeful.

[Mr SQL]

It would be awesome to write a game with a retro setup like that Agree Apples were comfortable to type on.

 

I guess you coded each bank as a seperate assembly file?

 

Awesome that you wrote your own Assembler at AdventureVision; you used this instant Assembler afterwards at Atari in lieu of theirs or did you use both sets of tools?

[+Gemintronic]

I use and iPad with DOSBOX for some developing. I wonder if most of your setup could be duplicated with an emulatior. Not entirely sure what a frob card is.