History and Rationale

History and Rationale

Soon after Tandy discontinued the CoCo 3, Color Computer enthusiasts and vendors set out to pick up where Tandy left off, continuing development along trajectories implied by previous CoCo generations. The two main contenders for the unofficial title of "CoCo 4" were the TC-9 Tomcat(optionally in combination with the TC-70) from Frank Hogg Labs, and the Multi Media / 1 (MM/1) from Interactive Media Systems, and later Blackhawk Enterprises. Each achieved some limited measure of success in the CoCo marketplace, but both were plagued with technical difficulties resulting in shipping delays and low availability, and were hindered by the limited resources available to the small businesses that promoted them. At the same time, all the non-IBM-compatible systems were disappearing, with the exception of Apple, as the age of the Microsoft-dominated commodity PC dawned. But more than anything else, one factor served to fragment what remained of the CoCo marketplace -- the lack of standardization between the competing systems.

The TC-9 took the more conservative approach, maintaining some degree of hardware and binary software compatibility with the CoCo 3, while the MM/1 maintained partial software compatibility only at the level of source code for OS-9 programs. The TC-9 was an advancement over the CoCo 3, but not sufficiently more powerful as to convince the majority of CoCo 3 owners to spend the money on it rather than one of the even more powerful M68K machines like the MM/1, FHL's own TC-70, or a generic x86 PC. The 68000-based CoCo 4 candidates, while much more powerful than the 6x09-based systems, were not fast enough for practical software emulation of the CoCo, which meant that those who wanted to run CoCo software that hadn't been ported to OS-9/68K would still need to keep a CoCo or TC-9 running alongside the new system. To further complicate the issue, other upgrade options, such as an HD6309 CPU upgrade, and Chris Burke's aborted Rocket design, made the tough decision of where to put a programmer's software development time even more difficult. Freedom of choice is a good thing, but schizophrenia in a marketplace as small as the CoCo's is disastrous.

So the fourth generation of CoCo development can be characterized by a lack of standardization, diverging technologies, and the encroaching influence of the commodity PC industry.

Now, two decades out from the introduction of the CoCo 3, the domination of the commodity PC is all but complete. There is no major computer manufacturer using anything but x86 architecture, and even Apple has dropped "Computer" from its name. Surely everyone who still owns or once owned a Color Computer now has a PC or a Mac, and likely has had more than one. Commodity hardware far more powerful than the CoCo is cheap or free. There is simply no way an updated CoCo design could compete with modern PCs on their own terms. But this fact can actually be liberating. Since nobody needs the CoCo to compete with the PC on a technical level, it removes a major constraint on the design of an updated CoCo. The money someone might spend on new CoCo hardware is coming out of the hobby budget, or the entertainment budget, not out of the hard-nosed pragmatic PC application budget. If the CoCo can no longer be considered a serious contender as a general purpose PC, there's no need to try and justify it as such. It can be what it will be.

So if the CoCo can't be seen as a realistic general purpose personal computer anymore, why still use it? Why update it? What are the qualities of the CoCo that keep people playing with it, working on it, discussing it, and celebrating it decades into its obsolescence? Though in many ways modern computers are indisputably superior, these advances have often come at the expense of other virtues.

Reasons to use a CoCo in the 21st century:

1. Nostalgia and Retro-Cool /p 2. Software Simplicity -- easy to program 3. Hardware Simplicity -- easy to interface 4. Educational Accessibility -- easy to understand. Provides a great platform for learning 5. Open Frontier of Opportunity -- Not everything has already been done on a CoCo. You can be a big fish in a small pond if you write a ground-breaking CoCo program. You can still be the first person to write a web browser for the CoCo. 6. Aesthetic Considerations -- There's something compelling about the feel of the CoCo. The rather cartoonish look of recent mainstream graphical user interfaces goes to show that garish colors and large typefaces aren't always a bad thing. More seriously, the CoCo has an immediacy and responsiveness that is lacking in modern computing, with its virtual memory and long boot-up times. 7. Robustness. If the CoCo loses power unexpectedly, the likelihood of disk corruption is quite low. It would only be an issue if the power loss happened during actual I/O operations. If you've saved your work, you can just switch it off. With modern computers, just powering down the system can be a chore. The user becomes a grovelling supplicant. With the CoCo, the user is master. Just save your work and pull the plug. And having BASIC (and potentially NitrOS-9) in nonvolitile memory dramatically reduces the risk of permanent operating system corruption. 8. Historical Preservation

With no further need to compete against the PC, CoCo development can (perhaps inevitably must) instead leverage the power of interconnected commodity computers. Hardware, software, and content development can all be facilitated by harnessing the capabilities of modern PCs to serve the needs of the CoCo community. The fourth generation of CoCo development was hindered by competition from generic PCs. The fifth generation will be made possible by collaboration conducted over an Internet infrastructure built on generic PCs. And while the fourth generation of CoCo development was plagued by a proliferation of incompatible systems, the fifth can be aided by community-developed standards for interoperability.

We are now living in an era when PCs can easily emulate simple systems like the CoCo at speeds much faster than the actual hardware, and hobbyists can implement 8-bit computers on FPGAs. Software emulation of the CoCo is here today, and FPGA implementations of CoCo functionality is all but inevitable. The only remaining questions are whether to enhance the CoCo's design beyond the capabilities of the CoCo 3 circa 1986, and if so, in what way. There are many reasons to emulate a CoCo in software, and many reasons to want to implement a new hardware design. But this time there are no excuses for not remembering the lessons of the fourth generation of CoCo development. Certainly now, if never before, the base of CoCo users and developers is too small to be split up into rival camps. A shared specification for a fifth generation CoCo compatible is essential. The talents and resources of no willing CoCo enthusiasts can be rejected out of hand because of their commitment to a rival, incompatible design.

By comparing the CoCo to the present computer paradigm, we can see not only how far we've come, but also what we've lost. This could just provide an inkling of possible future directions for computers in general. History moves in both straight lines and circles. Its helical nature encapsulates both linear progress and the repetition of cycles. The first great era of garage startups is long past. Companies have come and gone, and morphed into new forms. Industries have consolidated, with decidedly mixed results. Commoditization around a de-facto Wintel standard means interoperability and ease of use for the end user and the virus proliferator alike. But perhaps new technologies, such as software emulation, FPGAs, and personal CNC machines; and new methodologies, such as open-source style collaboration, can usher in a new era of hobbyist-driven innovation. Maybe an updated CoCo project can play some small role in such a future.

Joel Ewy