Message Board
  Atari Systems
  Game Scores
  Video Reviews
  Bonus Material
  Join TAT!


Reverse Engineering Rebuttal - The Atari Times

Reverse Engineering Rebuttal

Did Commodore Really Reverse Engineer the Atari?
by Charles F. Gray & Bob Yannes

November 17, 2005
I co-wrote an article back in April 2005 called "Reverse Engineering: Commodore Reverse Engineers Atari 8-bit Hardware" motivated by Carmel Andrew's email regarding the Commodore 8-bit line of computers. "Reverse Engineering" discusses our argument that Commodore engineers did in fact dissect and analyze Atari computers and game consoles to create Commodore's 8-bit line. Interestingly enough, I received an email from Bob Yannes, the Commodore Engineer responsible for the famous SID chip as well as co-founder of the synthesizer company Ensoniq, which makes a very strong argument against the aforementioned article. Little did I expect an email from the creator of the SID chip, but as with the email from Carmel Andrews, I will set out to explore what Carmel and I have formerly written by using Bob Yannes' email contained below:

Regarding the development of the Commodore VIC-20 and C64 with respect to Atari:

While the quotes attributed to me may be accurate, they are taken out of context and there is little accuracy to this account--it is based on a rather biased interpretation of history.

Although we were indeed trying to outperform the Atari line, as we felt they represented some of the best graphics and sound that were available up to that point, we did not reverse engineer the Atari products. There was never any intention whatsoever to hire Atari (or ex-Atari) engineers and the hardware designs are different on fundamental levels. We did copy their external connectors/pin-outs so that we could take advantage of various peripherals that were already available for the Atari line.

Bob Yannes states that Commodore had no intention of hiring ex-Atari engineers. Simply put, Jack Tramiel was a very shrewd businessman and would know the value of hiring ex-Atari engineers who would have both the experience and the expertise to bring Commodore to the next level. But Jack also knew that he would be able to get technical information from these same engineers whether they were conscious of it or not, as they themselves would use what they learned at Atari towards future Commodore designs. The thing is Atari ex-engineers were only hired during the Amiga era of Commodore, which was long after the C=64 was out and Bob Yannes and his fellow engineers left Commodore. But during the VIC-20 and C=64 era of Commodore, if Commodore could copy Atari's connectors/pin-outs, then what else could they copy?

There was no influence from the Atari VCS at all, which was a more primitive design than the 6560 VIC chip (designed in 1977/78) used in the VIC-20. The 6560 used transparent DMA to refresh the screen and did not require continuous processor-intensive intervention--this is a fundamentally different approach. The VIC could easily display both bit-mapped and character graphics for detailed color displays that were beyond the capability of the VCS. On the other hand, the VIC did not offer "player/missile graphics" or collision detection.

Yannes makes a very good point. Why would Commodore copy an Atari 2600. Simply put, they wouldn't. Carmel's argument that the Atari 2600 known as the Atari VCS was used in Commodore's 8-bit line struck me as odd. The VIC-20's graphics were more advanced than the Atari VCS, so it would have no advantage for Commodore to dissect an inferior design.

It would be more accurate to say that the TI 9918 video chip was a functional inspiration for the C64 video chip rather than the Atari products. We did *evaluate* the features and capabilities of many competitive products, with an emphasis on the TI and Atari Home Computers, as well as the Mattel Intellivision, but that was strictly to help us define what features and capabilities we thought were desirable. We did not reverse engineer anyone's circuitry, nor did we copy them. Anyone who knows the register maps of the two chip-sets knows that the Commodore chips are in no way "clones" of Atari hardware--that is rather absurd.

Yannes mentions that Commodore evaluated the features and capabilities of competitors' products such as the TI, Atari's line of home computers, and the Intellivision. To me, this is a form of reverse engineering. No they didn't copy the technology but the learned what they could from it and used these ideas to further their development of the Commodore 64. The C=64 is not a clone of Atari hardware. Contradictory to my own opinions and Carmel Andrews' opinion, Yannes states:

I also have to take exception with your definition of reverse-engineering. Evaluating the features and capabilities of competitors products is not reverse engineering. Every company operating in a competitive environment does that. It's part of market research--what existing features do people like, what existing features can they do without and what new features do they want?

Reverse engineering involves determining *how* a feature works and how it is implemented. This is where the "reverse" comes from--instead of designing a circuit to achieve a particular result, you look at the result and works backwards to determine the circuitry. This is typically done because the people who do the reverse engineering don't understand how to do it themselves. At no time did we examine how Atari, TI or whoever, implemented their circuitry, nor did we copy their circuitry. Quite frankly, we didn't need to--it was obvious to anyone who had ever designed video hardware how you would accomplish these things. I would give TI the credit for the concept of a "sprite" (as well as the name), but appreciating that independently movable graphic objects would be useful is not reverse-engineering. We didn't look at their circuitry to see how they did it, because we knew what needed to be done: address generators, data buffers and video shift registers, all of which we had designed before. Our sprites were more powerful than TI's, as they allowed multiple resolutions, scalings and colors and we had more of them. TI got annoyed at us using the term "sprite" and there was a half-hearted attempt to rename our sprites to "MOBs" (for Movable Object Blocks), but ultimately the word "sprite" stuck and we usurped it from TI as our products succeeded and theirs failed.

Among other aspects of the circuitry, the color generator used in the C64 was unique enough to be granted a patent, as was the SID chip itself. The C64 did not have "player-missile graphics", which we felt were too primitive, instead we preferred the "Sprite" model pioneered by TI. Atari did not invent color bit-map graphics (the Cromemco Dazzler preceded even the Apple II in 1976), nor did they invent character displays or digital sound generators and I have to add that the SID chip is still generally recognized as a groundbreaking achievement in home computer sound capability. Putting the sound and video on separate chips was not a nod to Atari, it was practical reality. Chip geometries at the time did not permit both functions to be integrated on the same chip, the die size would have been too large to produce. Also the noise level would have increased in both the video and the audio having the circuits in microscopic proximity. If we could have put everything on one chip, we would have, as it would have been much cheaper (less packages, no redundant circuitry, less soldering, smaller circuit board, etc.)

Yannes furthers his argument that the separate chip design was practical reality. He is quite correct. Also the SID chip was and is a groundbreaking achievement in computer sound capability. The SID chip blew away the beeping and clicking sounds of the Apple II and the mono-timbral very basic sound of the Atari 2600. The SID chip is still used today in modern music and was in my opinion what made the Commodore 64 computer leaps and bounds ahead of its competition at Atari and Texas Instruments.

Finally, my involvement with Atari when I worked for Commodore was limited strictly to ROM design for the VCS long before the C64 and had nothing to do with Atari's Home Computer products.

As an engineer, I admire what Atari did, but as an engineer who worked (very hard, under intense time pressure) on the Commodore chips and systems, I prefer what we came up with and don't appreciate our efforts being mischaracterized.

"Reverse Engineering: Commodore Reverse Engineers Atari 8-bit Hardware" did not mischaracterize Commodore but it did provide a very one-sided argument that tended to inaccurately overemphasize Commodore's efforts and tactics under Jack Tramiel. Commodore did come up with the majority of their innovations themselves, but also were aided by examining their various competitors' products.

Simply stated, Carmel Andrews and Bob Yannes have very differing opinions regarding whether or not Commodore reverse engineered Atari designs, which is based on their opposing views of what constitutes Reverse Engineering. The big question is: "Who is correct?" and that is dependant on the true definition of Reverse Engineering and whether it was actually done at Commodore. What we have to remember that some opinions are quantifiable fact but other opinions cannot be proven as they too much rely on speculation gathered from deductive reasoning or by a certain viewpoint that may or may not be objective. One opinion is from a well versed retro-computer hobbyist and the other is from a Commodore engineer. Both arguments have their merit. We must not forget Jack Tramiel in this equation. Jack Tramiel knew exactly what he was doing, his goal was not to play fair, but to come out with the best computer for the lowest possible price. Yannes states, "Jack believed that business was war and you didn't win by competing effectively, you won by destroying your competition." To do this Tramiel would have to explore all options to make Commodore a success. In truth, Commodore did create a lot of original innovations, but like other computer companies at the time including Atari, they dissected the competition's products to better learn how to achieve their goals.

Reader Comments for Reverse Engineering Rebuttal

Reverse Engineering by Sergius on 2007-01-18 08:37:07
I don't believe that there is such ambiguity in the definition of Reverse Engineering. It's an engineering term, and every engineer knows what it means. It means taking something apart, and deduce from its inner workings how it does what it does. I could as well say, "to me, industrial espionage is reading the newspapers to know what the other company is doing." That doesn't make it a valid definition. I agree that it's difficult to determine whether it was done or not, it can only be inferred, but your definition of reverse engineer is so broad as to encompass... pretty much every single product in the planet. Did the inventors of decorative wallpaper reverse-engineer ceramic vases? I mean, they probably evaluated its capabilities in "making rooms look nice".
Add Comment
What is the greatest video game company of all time? (Hint: Atari.)