Sierra, Gaza etc.

[Maury Markowitz]

Hi guys,

 

I'm the author of the Wiki article on the Amy, and if you haven't seen it lately, I've added all sorts of changes, so enjoy!

 

I was wondering if there is *anything* concrete about any of the 8-bit replacement projects. What I have found, mostly through email conversations with some of the engineers, is that none of the machines really existed beyond mockups. One engineer is even unsure whether Sierra had finally decided to use the 286 or m68k.

 

The only think I've been able to actually document is the Amy, which was helped in no small part due to a lawsuit. I'm worried but hopeful that some sort of engineering documentation from either Sierra or Gaza (assuming they weren't the same thing) still exists in someone's closet.

 

So here's the names I've heard, and I'd love to hear anyone's input…

 

Gaza - 68k based machine with Silver and Gold GPU and Amy sound

Sierra - 286 or dual m68k machine (seems unlikely!), Rainbow GPU, Amy sound, Eva OS

Eva - concept for the Sierra OS, existed only in theory. Main designer was apparently considering Unix V port

Silver and Gold - GPU for Gaza? Zero information on this

Rainbow - GPU for Sierra, and possibly other projects

Amy - the only part to be released

 

I have seen mentions of a new GUI and a different OS as well, here in the forums.

[Lynxpro]

Landon Dyer was very dismissive of the Sierra and Gaza projects in his blog, as if they weren't really serious projects. He stated dual 68000s wouldn't have been economical in the years following when he worked on the original ST. And since he was aware of those previous projects, it is probably logical to assume the Tramiels became aware of them too which kinda answers some of our ponderings on here over the years as to why Atari Corp didn't just roll out those earlier designs instead of pumping out the [allegedly inferior] ST.

 

Curt has mentioned in other threads that there aren't any [known] surviving screen shots of the Snowcap GUI that was planned for those computers running atop BSD [and not UNIX].

 

Can't wait to see the updates to the AMY wiki. I think I can say a lot of us would like to know why the Tramiels were unable to get the chip to work and on what legal grounds they were able to stop the company they sold it to from later bringing it to market...

 

Damn it, we want a working version of that chip to debut, even in FPGA form!

[Maury Markowitz]

Landon Dyer was very dismissive of the Sierra and Gaza projects in his blog,

 

Ohh, do you have a URL for this? Google isn't turning anything up of interest.

 

The dual-m68k is likely a canard. Any such device would have *serious* bus contention issues. You could do it, but the two CPUs would have to be isolated, and likely on their own RAM/bus. So maybe a CPU dedicated to the graphics or something, but that seems extremely unlikely given the costs. It wasn't until the 020 that multiprocessing became practical, and even then I know of very few machines that actually attempted it.

 

JerryM sent me a few things that suggest that the Sierra was definitely single-CPU, but the decision had not been made as to which one - 286 and m68k were both being considered. Curt has a block diagram showing a single CPU machine, with the NS32016 in it.

 

Curt has mentioned in other threads that there aren't any [known] surviving screen shots of the Snowcap GUI that was planned for those computers running atop BSD [and not UNIX].

 

According to JackP, the entire OS was just a paper project when it died. Unix was being "discussed" but it seems very little progress had been carried out.

 

The OS team are the only ones I haven't tracked down so far.

[Lynxpro]

I don't have a direct link to Landon's comments while typing this from my iPhone but he made it in the comments section in one of his blog posts discussing working on the ST - GEM specifically - and dealing with the higher-than-mighty folks at Digital Research. He briefly talked about some programming tricks him and some other former Atari Inc'ers knew from game programming that stumped the DRI staff. In the comments, someone asked him about why Atari Corp had plowed ahead with the ST if the Gaza/Sierra were completed in the lab and they had access to them.

 

It had been speculated by some on this board that some of those projects were"liberated" along with other goodies during the Tramiel acquisition of Atari Consumer since it had been said some employees were so cheesed off they brought trailers to the facilities and loaded the stuff up they had worked on so Jack and company wouldn't get their hands on the stuff unless they stayed employed.

 

If the AMY was as cheap to produce according to that Atari Inc email from 1984, it should've been used in every Atari project going forward...the XE line, the ST, and 7800 game carts. What a pity. It's also interesting to me how the later version of AMY had half the number of oscillators as the original Inc version but was supposedly more powerful.

[Lynxpro]

It's also funny how both Inc and Corp messed around with NatSemi CPUs before deciding they sucked.

[Maury Markowitz]

I suspect that the problem with Amy was that it…

 

1) didn't do everything the POKEY did, so you still needed one of those

2) wasn't ready in time for the rushed production of the ST

 

Sad, in the end, as the Atari line's sound performance suffered in comparison to its competition throughout this period.