Lew Eggebrecht, Vice President of Engineering at Commodore.

Interviewed by Mike Nelson for Amiga User International at the January 1993 Devcon in Orlando, Florida

Background

"These are all low end machines - all of my career has been spent in smaller systems rather than mainframes. We started a small test group to investigate very low cost products such as the IBM DisplayWriter and the first IBM machine with a BASIC interpret er. The outcome of all this was, of course, the IBM PC for which I was the design and project team leader. We did some further work on the architecture, some of which was implemented and some of which was not. I then went on to work on 3270 terminals for mainframes before leaving IBM to work for Franklin Computers who developed Apple compatible products. The big problem here was that they were unable to licence the software from Apple so they went out of business."

Lew moved out of hardware design into a software consultancy that specialised in telecommunications programming, doing projects for the likes of AT&T, British Telecom etc.

"We did a lot of packet switching stuff for X25 systems, and we developed the industry's first ISDN package on a PC so we could support the new digital communications systems that are beginning to appear in Europe and the US. I left them to start my own c ommunications company which I still own, although my Commodore commitments mean my family run it now."

How did you get involved with Commodore?

"Two years ago I started consulting with Commodore to help them develop their next generation PC compatible product line, so my initial work was not on Amiga at all. We're still one of the major suppliers in Europe of PCs but it's not as profitable as Ami ga though! Then, about a year ago, I began to work on the next generation of the Amiga's chipset."

I asked Lew what made him jump from the familiar environment of PCs to the radically different Amiga:

"Well, I'd worked on Apple, RISC, IBM and all that's left is Amiga - I knew nothing about it and it's a real challenge. The architecture is really interesting and has great promise. It's poorly viewed in a lot of corporate applications and we're trying to change that and generate a lot of interest in Amiga. This is one of the reasons why I'm going out and speaking to people, being more open with future strategies and directions to prove to people we are a viable vendor and we do know what we are doing. Th at's been a little lacking previously.

Obviously marketing is not my area but I try to promote the Amiga from an engineering point of view: the idea that the Amiga is a viable platform and we are doing interesting new things, and also that it's something new developers ought to consider."

This does go very much against the grain of Commodore's usual way of doing business. How much of this is your own personal influence?

"I think a lot of high level executives are uncomfortable with the technology or understanding and explaining it to others. I'm able to do that as I'm an engineer - I design things and do day to day engineering work so it's much easier for me. It's not th at Commodore didn't want to be open, it was just difficult with the management they have had in the past.

We're trying to demonstrate credibility and an ability to produce new products. We have goals. I think the fact we have released two new systems and other products inside a year demonstrates that Engineering is not asleep and can product high quality stuf f at a high rate. That's the story I'm trying to tell."

I pointed out to Lew that the Amiga group, both hardware and software, is a fraction of the size of the Windows crew and they are streets ahead in what they are producing.

"Very much so. In many ways the Amiga is just like those early days of the IBM PC when we were a small group of 12 people. It shouldn't surprise people that small groups do well - even in the early days of Microsoft, Bill Gates and one other guy did all t he work."

I've always wondered why IBM went for the Intel 8088 processor rather than the 68000 of the Amiga. Can you enlighten us?

"At that time there was the choice and one of the things that attracted us to the chip was the pricing we were able to get. Also we felt it was much easier to get software runnning on the Intel architecture rather than on the Motorola - at that time there were no native compilers so everything had to be done on an IBM mainframe and there was no development environment for the 68000. Also Intel offered a package that would translate 8080 or Z80 code so that gave us the opportunity to move some applications over quickly. This meant that there was a body of software that worked at announcement and we couldn't do that with a 68000. It really had nothing to do with the technical attributes of the chips it was more a business decision. Most programmers would pr efer the linear addressing of the 68000 which you now have with the Intels."

What sort of state was the Amiga in when you took over?

"Well just over a year ago, AA was sort of languishing. The design was done but there was a lot of bugs. We were too conservative in trying to put it into a product and flush things out so we decided we wanted a AA product by Christmas '92 and we achieved that. We put together a little task force, did a final run of the chips and got everything right and the products came out. The engineers were happy to have a project to work on so that lack of direction really was overcome. We put more focus on doing th ings quicker and better. We have some very free thinking engineers in both hardware and software - very very creative, and if you don't give them specific goals they'll just continue to develop and develop. The process of converting a design to a product is something Commodore has always had difficulty with.

We're stopping all that. For instance we used to do 4 or 5 revisions of a chip but now it's 1 or 2 at the most. We like to get it right first time and we now have a lot of powerful in-house tools and simulations of chips to help us do that. We're also usi ng a lot more industry-standard chips rather than unique things - if you leave engineers alone they'll re-invent the wheel every time and we can't afford to do that."

Tell me about AAA - it's been worked on since 1989?

"Yes, we worked on it from an architectural point of view for a long time but it's only been serious for about a year. It was obvious that AAA was not going to meet our cost targets for the mid to low end systems. We wanted to continue that development an d we also had to have an enhancement quickly so, AA was the solution to that problem. It would have been nice to have AAA at the same time as AA but we just couldn't get there."

What about this new AA+?

"AA+ will be a more profitable version of AA with all the things we wished we'd got in but didn't have time. We have a list of all the problems we currently have at the low end. The serial port, we can't read high density floppies, there isn't enough band width to do 72 Hz screens plus there are no chunky pixel modes for rendering. We listed all those and said "OK let's go out and fix them as quickly as we can", so AA+ is an extension, not radically new architecture. We're doing the best that we can, takin g advantage of advances in technology, significantly reducing the cost and that's the goal."

Where does that leave people who have bought A1200s now?

"It's going to be 1994 before you see any product. I don't believe there'll be any easy way to upgrade because of the packaging of the chips - they are surface mounted. The memory timing and interfaces are dramatically different as we're using a method ca lled split cycling to do two cycles at once. To get the video out faster we're bursting out four 32-bit words in one memory cycle so you can't upgrade a 1200. It's the same all over though - you can upgrade a 286 to a 386 if you change the motherboard -th at's progress. We'd love to ensure that no one was ever made obsolete but that's just not practical - you spend so much time being backwards compatible that it gets in the way of progress although compatibility is a major design target.

There is a limited amount of software available for the Amiga and you don't want to make that smaller. We want to ensure that if the developer does something legal, it will work on the next generation of the hardware. You may not be able to take advantage s of new features, but you aren't obsolete. Unfortunately software is sometimes written to be timing critical or uses a feature that we didn't know existed and got designed out, so that leads to problems. We spend a lot of time and money on compatibility testing."

Games programmers tend to be quite cavalier in their work, working to a shelf life of their product of only 6-9 weeks. What can be done to help them?

"We do a lot of testing, but programmers don't use operating system calls correctly, write directly to the hardware and have timing-sensitive code - changing the speed of the processor will make many games break. AAA has a lot of compatibility built in bu t we aren't 100% successful. We give out Beta systems early so if we can accommodate people we do. There's some pretty wacky programming going on out there!"

Where do you see the Amiga in a few years time?

"Clearly we will continue with the living-room type of box. We learned a lot of things from CDTV - where our best price point is, how important the quality of the software is and the fact that running Amiga software is important. Most of our sales come fr om applications where it is sold as a computer not a CDTV. We understand that we need Full Motion Video capability on the system and we are working towards all those goals. Getting AA out was important and now we have time to look at upgrading CDTV and al so looking at other price points. We can't do everything at once so we have to do what gives us the best return on our investments. Consistent high quality products are the most important right now but we will look at expanding the product line both above and below the current machines. Commodore makes a lot of its revenue from low end products so we pay particular attention to that area."

What about the long term?

"We want to see a complete family of products from the consumer level up to professional workstations. We have limited resources so we need to focus on certain areas. We hope that Amigas will become a standard in multi-media so it is important that in the future we run Windows NT and UNIX. I suspect that in five years time the Amiga will be RISC based. We aim to be the leaders in graphics, at least from a cost/performance standpoint although you never know, some of the things we are doing will allow us to bring the Silicon Graphics capabilities down to the desktop, and that's our plan."

It's rather ironic that you created the PC and along with Microsoft MS-DOS and Windows, all the bias against the Amiga. How can you overcome this so the Amiga is accepted as a viable machine for professional use?

"We can get to the point where the processor is not so important and the user can pick whether he wants to use AmigaDOS or NT, or PC graphics versus Amiga graphics at will. Once you get a person used to using an Amiga he loves it. That's always been the p roblem. Attracting developers to write software is another goal and that's hard because of the financial side of things. We do sell Bridgeboards - a 486 board is available today for the A4000, although we won't be doing one ourselves we'll be encouraging our third party developers to market them. We forsee very low cost solutions in the A1200 right the way up to Pentium products for the A4000, but again with the resources I have I'd rather encourage third parties to do these things. It's sort of off the s helf technology which we don't get much out of, so we may do joint ventures with people. In fact, one of the first things I did when I started at Commodore was move people from the PC division to Amiga, and we procure most of our DOS machines from Taiwane se vendors like other people (including IBM). I'd much rather use those engineers on Amiga projects than on PCs."

How many people do actually work on Amiga?

"Our total workforce on Amiga including CATS is about 175 people - as many as there have ever been."

What has impressed you most about third party products for the Amiga?

"Well from a technical point the NewTek stuff is very exciting. All the video stuff is impressive and feeds on the strengths of the Amiga. We are seeing the Amiga finding its way into very large vertical markets like information kiosks, video presentation s that are OS-independent. You really don't have to be PC-compatible for that. Performance, price and a good development environment is what counts and this is quite a large market."

Many people see the Amiga in much the same position with regard to video as the Mac was with DTP. Do you go along with this?

"Very much so. We have continued to preserve the NTSC/PAL capability as it is key to the system's popularity. There is no doubt about why the Amiga is unique and that is its interfacing with video."

What do developers need to do to break into these "vertical markets" and convince people the Amiga is the machine to use?

"In vertical applications, the system is hidden from the user - he does not make that decision for it made by the developers and when that is the case, we win because he understands the technology and its capabilities. It's much easier for the Amiga to go into this situation as there isn't someone in the board room saying "Why isn't this thing Apple or PC compatible?". He doesn't even realise there's a computer involved. You are selling an idea or concept rather than hardware - it works very effectively."

Are Amiga developers prepared for this?

"We know of a number of Amiga developers who are going after vertical opportunities. I think we're setting an example with our system software in terms of stability and compatibility, then performance. The new releases are much more stable. It's also much easier to migrate. That quality costs and while you can have brilliant people doing brilliant things, there's two sides to the product - the fun side and the hard work of making it stable."

Will Amiga software go up in price to be on a par with PC products?

"Not necessarily. One of Microsoft's pricing strategies is to use upgrades where once you buy into a product say for $600 and register it, to get to the next generation is only $100, and this has forced the overall pricing down. Access is the new Windows database introduced at $99. There is a strategy in the PC world to advertise at one price and sell at another! I think the Amiga will move in that direction."

What about the games market? It seems that the US games companies are moving away from the Amiga.

"The 1200 has changed things to a certain extent. We've had a number of companies enquiring about the Amiga but it is a problem we recognise. We have provided a low cost, high performance platform which should attract those companies back. Most our sales are in the games area, particularly in the UK. In fact, the UK has always been a bright spot for Commodore. Kelly Sumner has done a great job and the response to the 1200 has been amazing. We are building them under contract in Scotland and shipping them directly to the UK office - we can't make them fast enough."

Most of the games writers want to have 8 channels of sound. Is this going to happen at the low end?

"The current capability is four channels of 8-bit samples at 27 KHz and we forsee that most systems in the future will have CD capability. Most of the sound and music will come from this so it was not as important to put that technology in. Our long term strategy is to put the DSP in every system, obviously. That will be sound in and sound out and you can do pretty much whatever you like."

Does that mean that a AA+ machine will have a DSP?

"We can't make that decision right now - it's something we'll have to look at but in that time frame, even in the low end, every machine is likely to have a DSP. It's a cost thing - although the AT&T chip itself is only $20 to $30 or so. AT&T has a number of lower cost options, as well, that are designed more specifically to go on the motherboard. The problem with the present DSP design is that it has one serial channel and everything you attach to it has to be run through at that channel rate. I think th ey're looking at having four independent channels running at different clock rates, and with that kind of enhancement, DSP makes a lot of sense."

What about RISC processors and the Amiga?

"We are going down that avenue because that's the way the world is going. Even the Pentium is basically a RISC processor with a 386 core built into it. Eventually compilers will support superscalar design, executing two instructions with one clock cycle. Motorola architecture was actually much closer to a RISC architecture originally, given what we see from Motorola, the 68060 is realistically another year away and they are placing more emphasis on their PowerPC line, and rightly so. Therefore we have to understand what is going on out there and how to design a RISC system. RISC buys you the ability to port other operating systems like NT and UNIX and hence access to productivity software and a more professional environment. We can make Amiga video techno logy work with a RISC processor. It will be an Amiga though - you may be able to switch between an 040 and RISC processor, or emulate 68000 code. It's an extension of the Bridgeboard concept with two machines running at once."

Do you see Motorola stopping at the 68060?

"I don't know, but it's hard to see them justifying three processor projects, incuding the 88110, PowerPC. The 68000's will still continue to sell - in fact we are thinking of integrating the processor with the Amiga chips for the next generation machine. This gives a significant cost advantage and it could be faster, depending on the memory attached."

What's it like at Commodore?

"We're having fun at the moment - the engineers are enjoying things and we have a lot of internal discussions about what the Amiga should be."

Where is the main competition now?

"We do get squeezed with clone PCs at the top and Sega underneath, and also boxes like 3D0, although I'm not as concerned about this because of its price point. We are in that gap and below us there is no one with a stable operating system. We always have the option of cutting down but it's difficult for them to move up. PC clones are a valid concern as they're improving in capability although they're light years behind in understanding multi-media. There is a large enough gap and we're going to charge ou t ourselves."

What about the gap between the A1200 and A4000?

"We've addressed that with the A4000/030 which uses the 68EC030 chip and is much cheaper. There's no MMU but AmigaDOS doesn't need it. I can't tell you the price as I don't set them. The 040 is very expensive (hundreds of dollars), like the 486DX2 chip it 's a large proportion of the total cost. There's a lot of confusion about clock speeds and rating megahertz. It all depends on the memory outside the chip as well as the internal speed and system clock, so look at the benchmarks if you want to compare spe eds."

Why does the 1200 only have a 14 MHz processor?

"It's to do with the AA chips' architecture and chip memory. You don't get much more access to the computer's RAM by increasing the clockspeed as the chips have priority. In FAST RAM this is different, but the A1200 has 2 Mb of chip RAM. Adding a higher p erformance processor doesn't buy you a lot on the base machine - it just goes faster to wait more! Accelerator boards with their own RAM can go at any speed. Of course there is the old cost question - higher rate means more cost and we're talking dollars not cents."

Has the A2000 gone out of production?

"Almost. Pretty much the only avenue for this is the Toaster but Newtek is going to convert to the A4000, and they're really excited about the 030 version. Some vertical applications need the expansion slots but I don't know what it's life will be. Our in tent is naturally for everyone to move onto the 4000. The 4000T is close, probably in the summer sometime and yes, it's going to be more expensive because of the larger cabinet, power supply and more slots. Adding SCSI to the motherboard isn't so pricey, though."

How about that IDE drive in the A4000?

"Yeah, that. We always intended the AA chipset to span a family of machines and SCSI is not such a good idea at the low end. IDE is significantly cheaper than SCSI and the low end machines don't have the performance requirements. Unfortunately we introduc ed the high end first and really you do need the SCSI. Eventually we'll get it on the board."

Any message for the UK?

"Just keep buying those 1200s. We're really pleased with the response and the UK will continue to be a bright spot for us."

I hope from this interview to have shown what the current thinking inside Commodore Engineering is all about and personally, I am encouraged that Lew is able to announce the exciting developments currently under way at West Chester. This really is in star k contrast with the wall of silence that breeds rumours and untruths that usually surrounds the Amiga, although doubtless these will continue from some sections of the community. There is no need to propagate such useless information when there is so much factual stuff out there.

The Amiga is a machine that is going from strength to strength and Commodore has a great asset in Lew Eggebrecht, however he spells his name, and even the Padre himself, Jay Miner, seems to approve of the new approach. Let us hope the marketing people can live up to the great expectations of the engineers. Clearly the Amiga platform is going to diversify even further into the professional and consumer markets, and both Commodore and its developers must prepare for the awesome things to come.

Imagine if Lew is prepared to talk about all the AAA and AA+, RISC and such like, what is he keeping up his sleeve? A second report that was circulated on Usenet