The Case for PowerPC

Synopsis

The economic well being of the Amiga market demands a system that is not in direct competition with the Wintel world. Running on Intel hardware will jeopardise the viability of OS4 by placing it into direct competition with Windows. Intel derived hardware raises issues of support costs and financial return for the vendor.

Background

A replacement platform for the AGA Amiga. The AmigaOne was developed outside the confines of Amiga Inc., yet has been approved by Amiga Inc.

The Case for PPC

The rational for moving the next generation Amiga hardware to PowerPC architecture are primarily economic. Hyperion lacks the resources to test and support X^n combinations of motherboards, processors, and chipsets available today on the commodity x86 market. With Eyetech selecting a PowerPC design, there is a reference platform for OS4. Costs are reduced by narrowing the hardware possibilities. Software can be optimized for the reference platform. Executing old PPC enhanced software is easier. Selecting the PPC reduces transition costs and time to market. The PPC was in part designed by Motorola, maker of the 680X0 processors used in the Amiga. With limited time and money, this is important. One of the goals was to make a system where the user could run well written 68k software titles, and any software written for PPC accelerator cards. Much like the original Power Macs. The Motorola 68k series was chosen by the Amiga's original designers. Later CBM models continued to use the 68000, and in time the 020/030 and 040 versions. For some reason the 68008, 010, and 012 were not chosen. The 68008 was an 8 bit CPU, so the reason is simple. But the others may have been rejected for cost and performance reasons too. The entire system was designed around the 68000 and it's support chips. The software, especially file formats, were designed to take advantage of the CPU's architecture. A majour difference between the 68k and the x86 is the way the data is aligned. The Most Significant Bit and the Least Significant Bit are swapped. There are also issues on the way they access memory. In the case of the Amiga OS, memory access is important. If memory addresses don't align correctly, it is easy to crash the system. Motorola, like Intel, had their reasons for choosing their architecture's features.

The Supply Channel

With Eyetech as sole supplier of a reference PowerPC system for Hyperion's Amiga OS4, support costs are confined to a specific system, not any number of generic x86 combinations. To run OS4, you must purchase the system as offered by Eyetech. Eyetech can reduce cost by being the only purchaser of a production run, instead of a number of smaller, more costly jobs ordered by several suppliers. Going x86 would make it difficult for Eyetech and Hyperion to recover their R&D costs. Eyetech and Hyperion would have to select a specific x86 configuration, from many commodity x86 motherboard/processor combinations available. Eyetech would be unable to recover their costs through the sale of x86 systems, as they are commodities that can be purchased anywhere. Designing their own motherboard would be uneconomic. Another issue is will the selected motherboard be available in a year? Dealers too, would be forced to compete on the razor thin margins of the PC industry, and any extra OS4 support issues could make that course unattractive. Since users could substitute commodity x86 systems for Eyetech's system, there will be a number of users that refuse to purchase 'recommended' components, then express complaints when OS4 wonÕt function. Hyperion cannot afford to support users whose problems lie in their choice of hardware. Commodity x86 vendors have no interest in supporting a non MS operating system. If the user purchased the recommended commodity hardware, there is the matter of various components that may have compatibility issues. A tendency in the user community is to tolerate the shortcomings of Windows, yet blame the Amiga for any failings, regardless of their source. If Eyetech created a customized x86 board expressly for Amiga OS4, the dealer network would be placed in the position of selling a system that contains, in the mind of a potential purchaser, the same generic commodity components that he could purchase elsewhere, possibly at a lower cost. The average customer will not see any difference between an official AmigaOne and a generic pc. Additionally, some individuals could (and will) make allegations that Eyetech is charging a premium for a commodity product. There are people out there referring to the Amiga One as a "Teron", in spite of the fact that Teron's version is about four times the price. If Eyetech sold a customized x86 board, they'd make the same claim.

The Retail Channel

Retailers are in business to make a profit from their sales. They chose their products on offer based on the potential to make a profit selling them. Products with low margins, excessive support, or increased sales costs are not in line with the goals of any retailer. Computer retailers live and die on razor thin margins. Any increased costs related to the sale or support of a product can eliminate their margins. They also compete with web-based retailers, some of whom sell their low end product as a loss leader. Much like auto industry, money is made by up selling the product. It is more advantageous to sell a PC that can be profitably upgraded with a number of low-cost options. Since they build the system on site, any extra assembly costs are offset by the increased margins. Due to limited support within OS4 for many peripherals, dealers are restricted in the potential for additional sales with an A1. A printer may require an extra cost software package, for example. The limited OS support of graphics and sound cards, determines opportunities for cost reductions or increased margins that can be achieved through substitutions. As the success of the Amiga One is directly related to the success of the retailer channel, it is very important to offer a product that doesnÕt compete with the Intel architecture directly. With PowerPC, the dealer has a product that can be sold at a price unrelated to the commodity x86. In turn, the dealer isnÕt competing for a sale with every other dealer on the block. Potential customers cannot get a quote, then walk next door for another. The PowerPC AmigaOne offers dealers a comparable product, which can be sold at a better margin, to more knowledgeable clients, while avoiding the problems typical of commodity PC retailing.

Windows

PCs run windows by default. They are designed, manufactured and tested with that in mind. Microsoft sets specifications, but the manufacturer is free to implement them in any way he may see fit. The fact they run Linux is a bonus. If the system can be configured to dual boot, the user can install Windows and OS4. The user will be tempted to boot Windows to access a website, or run any number of network applications that are not available for OS4. If the user can run a browser such as Internet Explorer, what incentive is there to purchase IBrowse? The Amiga OS would soon suffer the same fate as the BeOS. Users would install it, play with it, say it's neat, and then they will reboot into Windows and forget about the Amiga OS. Users will also be tempted by Windows software, for many reasons. Users will purchase Windows titles because: They can buy them everywhere, today. They will not have to wait six to 18 months for titles ported to OS4, if that happens. If a ported title becomes available for OS4, it enters into price competition with the windows version. At the time it becomes available, the Windows version may be discounted, making any price difference appreciably apparent. They can play now, or wait, and if later they can buy an OS4 version, it will probably cost more. The same would apply for applications such as video editing, DVD authoring, etc. Cost is a majour factor when people make decisions. The possibility of users purchasing Windows titles makes the reward for porting to OS4 less attractive. Even with a reduced licence cost, sales volumes may not justify cost, or allow competitive retail pricing. Entering a price driven market will make it even more difficult for software houses porting to OS4. Unique, OS4 exclusive titles with high costs and low sales volume make it difficult to justify the effort. Retailers are not interested in setting aside valuable shelf space for slow moving, low volume, high cost titles. Exclusive or not. Software publishers will have difficulty competing for shelf space in that environment. The Amiga market will not survive unless it expands. By running on PowerPC, the spectre of competing against Windows on the same hardware is eliminated.

Piracy

By limiting access to OS4 by means of a PowerPC AmigaOne, there will be fewer unlicensed OS4 installs. If OS4 ran on commodity pc hardware, two issues would arise: 1 Large scale piracy, resulting in lost revenue to OS4's publisher. Problematic installs, using a questionable copy of OS4, in turn lead to a poor reputation for OS4 amongst the general computing community. 2 Compatibility issues within the hardware combination, for which OS4 takes the blame. This would hinder sales for new Amiga software, and damage the Amiga brand name. You want OS4, you must accept the hardware too. Otherwise, OS4 installations would soon vastly exceed OS4 sales.

The User Experience.

Tying OS4 to a reference platform benefits all the players. The OS and hardware can be optimized for performance. In turn, a stable, trouble-free system results. This allows a positive user experience. With commodity pc hardware, it would be very difficult to integrate hardware and software to the same degree. This would create an environment where the user experience could vary greatly, which in turn could result in negative impressions of the Amiga and OS4. If the user has spent a significant amount on his AmigaOne setup, he should expect no less than a positive experience with it. Assuming an x86 version, a significant number of OS4 users would not have a positive experience with their x86 AmigaOne, as they will have purchased commodity pc parts that will not be recommended, for monetary reasons. This in turn leads to support problems for Hyperion, and negative press on the internet.

Writen by: Michael "The Terminator" Rozeboom

Published: 2nd March 2005