The future of the Amiga


    (typed by Headhunter! (also called [-hhR!] !) from the Ruhrbasin!)



                   Europe is now a key to CBM's future.
     Lew Eggebrecht is well aware of that and is a virtually full time
                         Transatlantic traveller.

                  » ?: is AmigaUserInternational (AUI)) «
                  » !: is Lew Eggebrecht                «

?: How is AAA progressing - is it still the same as we heard in Devcon
   at the Developer's Conference.

!: Yes. The specifications for AAA have not changed from previously - in
   fact it would be very difficult indeed to change things at this late 
   stage. We are in bug fix mode only and we have mostly working chips at
   this time. There are some video modes that do not operate, there is a
   problem in the Colour Lookup Table and that's the first chip we are
   correcting so we can get on with testing the rest of the system. We
   already have very significant portions of the AAA system running. The
   turnaround time for manufacturing a new chip is around 6-8 weeks so 
   that's where we are at present.

?: What sort of machine will this go into?

!: Initially it will be a next generation 4000; we've designed the board
   - in fact that's the test board we are using. Assuming that the next 
   pass doesn't have any major problems, we should have it in a product
   by mid next year (1994). This will probably be an 040 based machine
   at a higher clock rate than the present system, either 33MHz or 40MHz.
   We'll definetely have VRAM instead of DRAM.

?: Is it going to be called the A5000 and what specification will be 
   possible?

!: Internally that's what we refer to it as, but that could change at the
   last minute. The CHIP memory can be up to 16MB and the machine will
   allow up to 64MB of RAM on the motherboard. 16Bit sound, 8 channels and
   100KHz sample rate for the sound.

?: What sort of software support is under development?

!: We're planning a major OS upgrade to support the new features. There
   will be a lot of trivial features which aren't worth the effort of direct
   support at present, but all the multimedia features like True Colour,
   chunky pixels, the new peripherals and full 24-bit graphics in 1024 by
   1024 pixels will be there. HAM modes are essentially programmable in
   software - virtually any mode will display it, but most people would
   prefer to see their images in True Colour. HAM is extended in a 
   consistent manner for animation purposes. There are other compressed
   video modes in the chipset which don't have the problems of HAM.

?: What about retargettable graphics?

!: Yes. That will be a priority as we have to do device drivers for the
   new modes, anyway.

?: What about the changes of staffing at Commodore Engineering?
   A lot of bad publicity has been generated by the lay-offs and 
   resignations of key people in development. How does this affect the
   evolution of the new machines? Is AmigaDOS safe from WindowsNT is the
   question many developers want answered.

!: No one is working on NT, so it's not an issue. Everyone wants the top
   productivity software on the Amiga but is no way that you're going to
   get Microsoft to port their stuff to the Amiga. In supporting NT we 
   will address this problem. It does not mean that AmigaDOS is going to
   go away - it's something we have to do. The beauty is we don't actually
   have to do anything to run NT, just choose the right RISC processor.
   If it's already supported by Microsoft, there's nothing for us to do.

?: Will that coincide with AmigaDOS in the same machine?

!: Yes, the intent is to be able to switch between one OS and the other.
   A Motorola processor and RISC processor would be present and you just
   switch between them. AmigaDOS will not be ported to RISC - it just
   wouldn't work as it isn't just a simple re-compile. There is no real
   need for both AmigaDOS and NT to be running together except for file I/O
   - it looks nice like with the bridgeboards, but isn't necessary.

?: So AmigaDOS is safe then?

!: Yes, very safe. In fact we are looking for a number of new homes for
   AmigaDOS in satellite and cable receivers, multimedia players. We are
   actively pursuing those areas with both the Amiga chipset and AmigaDOS
   because it's a much better solution than Modular Windows. It may not look
   like an Amiga but we are taking advantage of the very high performance
   multitasking operating system, a mature development environment and video
   capability.
   In future then, you may have an integrated box containing your TV, video,
   cable receiver with 500 channels and an Amiga running it all. Commodore
   are looking very much further forwards than most people give them credit
   for.

?: How about the 1200 in the USA?

!: We sell very few A1200s in the States - the pricing of PCs makes it very
   difficult indeed to sell them. It's much better in Europe where it's 
   competing against SEGA but that's OK as we're clearly superior to them.

?: The CD32 is very important to Commodore. How is it going?

!: At present we are just ramping up production. As far as marketing goes,
   we feelit's just a matter of getting people into the stores and they can
   compare us to the SegaCD and there is no competition. We're significantly
   better, so the lack of $25,000,000 marketing budget doesn't bother us as
   much as people may think it does.

?: How do you go about designing a new system like AAA?

!: The cips themselves are designed at Commodore by the VLSI group headed by
   Ted Lenthe and simulated on Sun workstations. Test data is also generated
   on these machines, but in order to see if the things actually work, you
   have to go out and build a chip. That is done through a major semiconductor
   supplier - we have been working very closely with HP (Hewlett Packard) and
   also with LSI. They build a prototype, which we get back and put it in a
   system. System level design is done by another group headed by Jeff 
   Franklin. Under Jeff are three separate departments, High End design (where
   the infamous Dave Haynie turns out his designs), another group for low end
   systems and the third does all the certifcation of systems, such as the
   reliability of a chip.

?: Where does the specification come from?

!: This is a key part of the process and variety of different interests have
   an input. It's a cooperative effort between all the groups and once the
   specification is known, the chip characteristics fall into place - it's
   kind of mechanical what happens from that point on. Specification is
   normally done by Jeff Porter, from my advanced technology group,
   Dr. Ed Hepler, the chief architect, and myself. I get heavily involved
   in this stage together with Ted Lenthe and head of system software, Allan
   Havemose.
   We have meetings, sometimes very wild meetings, to trash out the
   specifications. After that, it's work, but not really innovative work, to
   make the product happen.

?: What sort of timescale do these things have?

!: It's a gate array design, you can probably do the product in 9-12 months.
   If it's a fundamental chip re-design, it's around 18 months to two years.
   AAA is full design whereas CD32 was just an extra gate array. A gate array
   is like a customisable chip so rather than designing each transistor from
   the ground up, we just configure a standard array and wire them up. You can
   therefore make a new chip at what they call the "metal level" very quickly.
   Akiko in the CD32 does all the kinds of things, controlling the drive,
   system glue, chunky to planar conversions - almost anything.

?: What are the advantages in doing a custom drive as in the CD32?

!: The drive mechanism is a standard Sony but we did the control hardware.
   This means there a number of tricks to gain more speed out of the system
   and super low cost.

?: AAA is now working, but what next?

!: It has to be thoroughly tested which actually takes quite a lot of effort
   and is really important. The last thing you want to do is commit to second
   pass silicon and still have a bug as this alters your schedule dramatically.
   We had this problem with the Buster chip in early A4000s. When it was
   designed we had no Zorro III boards to try it out and it slipped through.
   Our track record on revisions of chips has dramatically improved. 
   Previously it would be usual to do four or five revs of a chip, but now
   it's down to two. Often we are getting it right first time which is a 
   great credit to the designers. We are being much more careful and paying
   more attention to the simulations.

?: How about the testing procedure?

!: There are two considerations here, firstly the manufacture process and also
   the design. We can generate programs in an Automatic Test Generator which
   will hammer the circuitry of the chip and look for problems, but this will
   also test bugs as OK. To find design errors you have to do system level
   simulation, writing test vectors to simulate what the chip does and this
   is much more difficult as the programmer has to know exactly what the chip
   does and the system it fits into. Imagine the complexity of testing a chip
   with 100,000 gates. You could go on forever testing and you need to draw a
   line somewhere and just go for it.

?: Where does AAA stand in all this and what about the software?

!: We're doing system level simulation now and the software people are key to
   this phase. One of the best tests is to try and bring up the operating
   system of this thing. System people are writing software to test new modes.

?: Where does manufacturing happen?

!: We have a state of the art production plant in the Phillipines, with brand
   new equipment. There is a well-educated workforce that costs about a third
   of what we'd have to pay in the US. This time of year is difficult as we're
   working flat out; sometimes we manufacture at other plants, such as 
   Scotland, but it takes three months to get a line up and running. There are
   procurement offices all over world for buying parts - Germany, US, Japan,
   and Taiwan so we can aim for the cheapest prices.

?: So what do the Advanced Technology people do as their contribution to the
   specification is finished?

!: We are starting to design the next generation - these things overlap and
   this is necessary if the two year cycle between products is to be
   maintained. We are concluding the specification. Our target is to have new
   technology every twelve or eighteen months - that's the same as the PC. We
   usually offer a tradeup deal which is much more than PC-users get.

?: Where do you see the Amiga in 5 years time?

!: The number one multimedia computer with software support across the board
   from AmigaDOS, Unix, WindowsNT and maybe even Mac eventually, but I don't
   think that Mac can survive as well as the PC. We want to be the one
   platform you can do anything you like on and have superior audio and video
   capabilities. There's no technical reason why we can't do that.

?: How do you see the other machines at present?

!: I wouldn't want to underrate the competition and the PC has made great
   strides recently. A recent poll in the US suggested that people found 
   Windows 3.1 easier to use than System 7 on the Mac, which is surprising.
   There are very high performance video boards available for the PC nowadays.
   PCs were locked into this 100% compatibility thing for many years. They had
   the same problem we now have, but they solved it earlier than us - if you
   write to the WindowsAPI, other boards are automatically supported,
   retagettable graphics... We have a leadership position in multimedia and we
   want to maintain it, but you can't underestimate the PC. It's important we
   get to the point where we can run PC software as that gives us credibility
   in a lot of corporate markets which we currently don't have.

?: Have you had much success in converting games companies back to the Amiga
   and are you going to promote CD32?

!: We've just started that and I'm not too sure how it's going yet. The
   strategy for CD32 is to go slowly at CES. We don't have a lot of money to
   spend in the US and there is no need; we can sell all we make in Europe, but
   we will overcome that and there will be a point when we will want to be in 
   the US, but we don't have the finances to support it prperly.
   We are actively looking for partners, something Commodore hasn't done in the
   past.
   A first example of that is the deal between Commodore and ASA (The number one
   manufacturer of IBMs in Taiwan). They will use our brand name to sell their
   computers, so we don't have to do anything except sit back and collect money
   for them using our name. It's a useful business model for addressing markets
   and geographical areas we can't otherwise reach.
   The same is true in technology when we've always gone it alone, but we're
   working with a number of other companies like HP with respect to the new
   chipset. That's the kind of thinking that's gone on in the last year - I
   just hope we can do things fast enough.

?: Much has been said and rumoured about Commodore's legendary changes in key
   employess. Who's Who at Engineering?

!: I'm not part of the US company I report to Mr. Ali, president of Commodore
   International and there are a number of departments. Contrary to all the
   rumours, there's a lot of people at Engineering and there's some sort of
   structure to it. When I took over, I reorganised this from a mish mash into
   something more structured. It seems to work - although the company has had
   financial troubles, the products are coming out quicker than ever before.
   We now have roughly around 100 people - we let 50 go in the last year, but
   most were in the support areas and we tried to keep the core technology posts
   as they are so important to the company. Engineering took much less of a hit
   than other departments. We've have some resignations, but I'm the only
   organisation within Commodore authorised to fill vacant posts at this time.

?: We keep hearing about resignations of key people like Peter Cherna
   (Intuition Guru), but how much does this affect things?

!: I understand the concerns of people. The industry in general is in a turmoil
   and the number one concern of people is job security, then "Is the work
   interesting" and thridly the money. I do everything I can to help and I 
   don't think any of our resignations have been unfriendly. I know about them
   long before they happen and if there's something we can do to help, we will.
   If not, we hire other people!

?: People are concerned because there is such a small core of programmers with
   in-depth experience of the Amiga and they see these experts drifting to
   other platforms.

!: There was this cult thing years ago about the Amiga. People used to think
   Amiga was so mysterious taht they couldn't learn it, but it's just not true.
   It only takes a good engineer maybe three or four months to come up to speed.
   When the Amiga was introduced, a lot of people didn't understand what it was
   about, but there's a lot of good poeple out there who do. This whole thing
   about there being a small core group who know Amiga is simply not true.
   Several have gone to our developers and even to competitors. We replaced
   the guy doing the Akiko chip and the new engineer just found some bugs which
   were corrected, so change isn't always a bad thing. I don't deny, it's a
   problem when key people leave, but I can deal with it. That's my job.

  Since Lew Eggebrecht has been doing the job, the effectiveness of CBM's
engineering has, in the view of most Commodore observers, made substantial
   strides and the speed of technical development notably accerlerated.
 Under this leadership CBM's excellent record of technological innovation
looks likely to go on keeping the Amiga as a major power in the computin world