Submitted by Prior to that epic event, however, there was another Amiga – a lesser-known member of the family most have never even heard of. Back in 1984/1985 Commodore created a few hundred “Development Edition” machines called the Amiga Development System. Sometimes, due to a very unique early design, they are also sometimes referred to as “Velvet” which was a name for a particular motherboard layout some had.
Commodore sent these computers to companies around the world in the hopes they would decide to support the new platform in the form of creating software and tools.
Thus, the Development System is a very unique machine most of which have been lost to the sands of time. Prior to this writing it was believed that only 5 Development Systems remained around the world.
Assuming that’s true, there are now six.
As indicated, this is an incredibly rare Amiga machine, so it’s probably the only time we ever get to see such a close and detailed look at it. The linked article contains a detailed video of the outside and inside of the machine as well.
Very cool find!
Very interesting! There is further info in the comments, such as about the difference between “Velvet” and “Zorro”. I wasn’t aware of some of this.
Zorro was the name of the expansion edge connector on the side of the A1000 and A500. The slots in the A2000 are actually called Zorro II. The Zorro expansion and Zorro II slots have a 16 bit data bus and 24 bit address bus, along with a bunch of control signals. The A3000 introduced the 32 bit (data and address bus) Zorro III slots. If you’ve ever looked at Zorro III, you’ll find that PCI looks A LOT like it. I’d called PCI the spiritual successor to Zorro III.
I could be wrong, but if I remember correctly, Zorro slots (all versions) were more or less just a standardized connection to the main CPU’s address and data bus (plus signaling), which is much closer in design to the PC’s ISA bus (Zorro II would be a lot like 16-bit ISA, Zorro III would be more like EISA). Again if I’m remembering properly, PCI put a standardized abstraction between the CPU and the slots/cards. This is why computers not based on x86 were able to fairly easily incorporate PCI slots. Adding ISA slots to a non-x86 system required either simulating the x86 bus, or throwing an x86 CPU into the mix and communicating between the x86 and non-x86 CPUs.
The signaling was based on an asynchronous bus, which the 680×0 supported, which is different from PCI, but Zorro II and III had the basic blocks of PCI built in: configuration space to allow boards to be identified, mapped into memory and/or io space, initialized, and load a driver (from disk or embedded).
Indeed.
But in the comments/updates, someone mentions that the earlier prototype machines were code-named “Velvet”, and the later ones (like the one in this article) were code-named “Zorro”. According to this comment, they kept the “Zorro” name for the slot because “no one suggested a better name”.
That was news to me.