The following text is from an audio tape furnished to me by Alan Crandall from the Redwood Empire Amiga Users Group (REAUG) in Santa Rosa, California. Alan taped this at the AmiWest`99 Amiga show in Sacramento. This was a conference call from the World of Amiga show in London, UK.
I've tried to type the text verbatim. However, there were some technical difficulties and I wasn't able to hear all of the statements, questions and answers. I'm sorry. Maybe our Amiga friends in the European Community or other areas of the media who successfully taped this talk can fill in the blanks or make the necessary corrections for us.
Naturally, the information about Amiga's Multimedia Convergance Computer (MCC) is no longer relevant as the new Amiga president Thomas J. Schmidts has canceled those plans. Information about AmigaSoft, AmigaObjects and the Amiga Operating Environment are supposedly still on the books and may be of interest here.
The text begins after x-president, James Collas' speech and a long period of applause. Amiga, International president, Petro Tschschenko is speaking.
"You all know how I'm thinking when I started in 1982. And then `85, this revolution stuff in my life started. And now, I'm very proud with my stuff, to take the opportunity to go with Jim, with our team, into the next revolution. So please, join us,join me, be a team, of our revolution."
"We have, from the US, some more people brought over from our team. It is Dr. Rick LeFaivre, Senior Vice President, R&D. Rick, here. As you know he has worked for several years for Apple. Correct?
Dr. LeFaivre answers, "Sorry." More applause is heard. Mr. Tschschenko continues, "And I think a lot of you know Dr. Alan Havemose. He worked with me in 1985, in the European [division] head for OS 3.1 and then he moved to the US. So Alan and Rick will tell you something. So please."
Dr. LeFaivre takes the podium. "Thank you. I want to say that I believe, I believe. Raise your hand if you have a Suit Coat on?" It sounds as though Dr. LeFaivreLafavre is taking off his coat. "How is that? OK, It's great to be here. I can't walk on that side. We're going to try to do a theme talk here. I'm going to start off with a little bit of over-view. The main thing we want to do here is answer some questions that have been asked."
"So let me first start by talking about some of the big megatrends that are happening. These are the things that are really driving the whole industry that you need to be aware of. This big convergence thing; we use the term Multimedia Convergance Computer because convergance really is happening."
"Ultimately a bit is a bit. The computer doesn't care whether that bit came from a video or audio, a picture, or whatever. So we've got to understand that convergance happens to be one of the chief drivers of the early part of the 21st century."
"The Internet is real. It's here. Internet collectivity is really happening. I think that the US is probably a little bit ahead right now. I think that the rest of the world will catch up very quickly, in Europe, in Singapore, places like that. We're getting to the point where every home is going to have a fat, broad-band pipe coming into it; just shipping bits down. That's very important, how you take advantage of that."
"Internet appliances; Jim talked a little bit about that. They are happening. It's going to be the next big wave of the 21st century. Portable software; we've been after this for years with computer scientists. How do you really get portability? I think things like Java, Jini and AmigaObjects are really going to drive a revolution in portable software."
"And this phenomenon of the Open Software movement. Linux is real. It's amazing, because we went out and talked to partners about what we're doing and talked about our choice of underlining OS kernal. So many people have said that we're going with Linux. They want an alternative to the evil empire out there. And Linux people are rallying with the Amiga."
"So there is a real dominance. This really is happening and you're going to be seeing lots and lots of announcements in the coming months as more and more people jump on this band wagon."
"Now Jim showed this slide. When I go into companies three months ago as CTO, and Jim unveiled this vision to me. In fact, it was kind of a funny little story. As he was recruiting me, we had several dinners and he'd begin to talk about his vision but I hadn't signed an NDA; and he'd say, 'and the next thing we're going to do is... No I can't tell you that.' and I'd come back for another dinner, he'd start to tell me more too, 'and then we're going to...No I can't tell you that.' So I had to join the company just to find out about the whole thing."
"As he unveiled it to me I thought, as CTO, Oh shit! There's a lot of stuff that has to be done to make this happen. I know that Alan had already been working a lot on these features, and he'll be telling you that in just a second, but there's a lot of stuff that has to happen. But luckily, we're at a very interesting time in terms of technology that are becoming available."
"There are a lot of interesting technologies out there that are coming up with technology devices. People like Sun and other companies that we can't name yet, etc. I've had to change those slides from our internal presentations in places. There's a lot of stuff out there, but people aren't putting it together yet. When we met Sun Microsystems to talk about Java, which is some really cool stuff there. They don't have a clue what to do with it. And as we, under NDA, unveiled our strategy with Sun, they said, 'Oh god, this is great! Please, we want to partner with you. Could we invest in you?'"
"So this is really, really, exciting, as we begin to put this together. Our job, of course, in this wonderful graphic, is to begin to utilize these technologies in interesting ways, and we're using the AmigaObject architecture as kind of the unifying architecture to put them all together. So you're going to be seeing AmigaObjects everywhere, levering technologies all over the place."
"So what are we doing? Well, we're building the Amiga Operating Environment; we'll be talking about that in a second, built on top of AmigaObjects; Alan's going to talk about that in a second. We're building one device called the Amiga Multimedia Convergence Computer, that actually comes in two forms: the integrated form in that beautiful case you saw out there and a board for more flexible configurations."
"On top of all that, we're building an information appliance environment which is portable, that Jim's said, that will work on top of anything out there that has a Java VM. That's the one requirement. And we're doing some reference designs, working with some partners, on a whole variety of different information appliances computers of one kind or another, portable computers, etc."
"Now this is the highest optical abstraction level of the Amiga Operating Environment, or software architecture and I'd like to have Alan say a little bit about what's really going on in the different levels of the architecture."
Dr. Havemose steps up to the mic and says, "I have to check this high tech devices out. I guess that I will be, later in this talk, but also tomorrow, giving a detailed actual session where I've got diagrams. So I'll give you a very brief overview."
"On the lowest level, we have the hardware platform, and as you can see, we haven't specified a processor, system, a disk sub-system, and any of that stuff, and quite frankly, it really doesn't matter. We're really building an architecture that is fairly independent of what ever busses that you pick or what ever 3-D graphics sub-system that you pick. The only thing that we're really interested in is the capabilities of the system and us tieing into it."
"The next level up, we come to the Operating System. In some sense, if you look at the whole picture, this actually indicates how significant to us, the decision to go with Linux is. Look at the total number of pixels working on it, Linux is just a tiny piece of it. The choice of Linux over QNX really doesn't effect it that much. The reason to go with Linux that we're working on and what Jim talked about, the fact is that when you look at Linux in the over-all architecture, it's a fairly minor detail but allows you freedom in a number of ways, and as you can see in the address, a number of other functions also. So you can easily add QNX or another Operating System or an Environment on top of Linux."
"The next level up is what we call the software interface level, and that is really the system software PARCing level. This is where the Java VM sits. This is where multimedia, the streaming devices will sit, in open GL in that level, computer graphics 3-D graphics, and what ever else you want to do, standard API, for instance the cabling, controlling APIs. It's all open to that level."
"And now we are up to what is the software livelier level. This is where it really happening. This is the guts of the Amiga Operating Environment. This is where the Java classes sit, this is where the organization of AmigaObjects sits. And they have to base themselves on to what ever the PARCing under it, such as the multimedia, Open GL and the Operating System. All that stuff is being built-in to the AmigaObjects. I think that you can see when you look at the software stack, it can now becomes transparent to anything that runs in the Amiga Environment, as long as they make themselves on AmigaObject, and the Java classes.
In many cases the Java classes are actually Java working in our AmigaObjects. All the software library or what ever will be the different instances of AmigaObjects. I think that you can now see that applications really don't care what Operating System they are on. They don't really care what particular implementation. The only thing that they care about is how AmigaObjects actually work and communicate. And then you can see that the environment you typically call the desktop and how you write your applications and that's followed by applications sitting on top of that. We'll follow that in detail in the seminars. We really don't have the time to cover that here."
Dr. LeFaivre moves up to the mic. "Just one comment from my perspective. One thing that you worry about as an architect or computer scientist when you see that is, Oh my God, there's a lot of different levels. Is this going to be inefficient? One of the key elements that Alan has been working on for months now is that its a very tight mapping. So you don't loose efficiency as you go down the different layers. We do think about things like that."
"In terms of the Amiga MCC, we're not going to be spending a lot of time talking about this. People have been speculating about the processing sub-system. Trust me, it's really cool. We'll answer questions about that at the appropriate time. All the hardware guys are here quickly trying to use their photographic memories to look at all of the different pieces."
"The main thing I'll tell you is, people have been sending us notes asking us, why are we using this versus that. We think that this might be better. Believe me, we are not going to purposely choose anything that's not the best thing that we can choose. So we're evaluating all kinds of things, working with the different partners, working with the top hardware companies in the industry, top graphics companies. Companies like BroadCom who has 95% of the cable modems of the industry now, and a whole bunch of people."
"The major guiding philosophy behind the architecture, is leveraged standards work makes sense. We don't want to be dumb and try to create something that's proprietary. On the other hand, get rid of as much of the PC baggage that we can, and so we're trying to move forward as much as we get through the thing. It's going to be a very exciting computer in its own right. But understand that this is only one piece of the over-all puzzle. Exactly what this particular device looks like as it comes out."
"Hardware products; we talked about the integrated MCC. We talked about the board, that's actually a generic board, I must say. So don't worry about it as apposed to the actual board. And then some reference designs for portable devices of one kind or another. So you'll be seeing a lot of this stuff come out."
"Now, what we really wanted to do in our session was to answer questions. We've had a few e-mails as Jim says over the past couple of weeks, since we released the technical white paper. We've boiled them down to the top ten questions. So my job is to be the questioner right now. Alan's job is to answer the questions. I will tell you right off the bat, that the first and last questions are by far, the most important, then there's eight in the middle that are, OK, you're interested but are not quite as important but we'll answer them anyway."
Question 1: "AmigaObjects. We've talked about AmigaObjects. It seems really key. What are they really?"
Answer: "I guess, kind of a history here would be appropriate. When I began as a designer of the AmigaObjects architecture, almost a year ago at this point, I thought of removing all barriers that I could think of, to the typical computer architecture. There's a couple of them I was really thinking about. One is that we tend to look at a computer as constrained by the boxes they're in. That is why Jim was talking about the AmigaObjects, networking is intrinsic. What that particularly means is that when a different object needs a resource, it doesn't know, it doesn't need to know, and it doesn't care, what defines the resource in the computer, in the box, that it's running on, or if it has to go out on the network and find it, and that's a very powerful concept. Everyone is talking about networking but I'm sure that all of you who has worked an Internet devices and particularly in PC space know that it's very unreliable. It's very difficult to find resources on the network."
"We have a thing on the Amiga architecture a way of transparently finding and using the network resources. And that really is the key to much of what we do. And secondly, multimedia, streaming mediatypes are absolutely essential. Any type of modern computing device will have a very high performance, real time streaming media (audio, video, MPeg, 3-D graphics) all the typical things that make the Amiga an Amiga, and all the intrinsic hidden AmigaObjects as well. On top of that, we have access, authentication, security controls, and encryption, all the stuff that it take to actually allow the Amiga to be trusted on the network both in terms of telling and convincing other people that an object has the right to ask a certain kind of service, but to all types of things in related to certain kinds of objects that survive, live and manage to come up on the network." "Finally, there's a piece of the design that I don't want everyone to be hung up on a programming language. As a matter of fact, I did the original implementation of this AmigaObjects in C++ as I was working. The fact is that Java really becomes more and more important and the architecture is super clean. I translated the whole thing to Java in less than one day. So that gives you an idea of how flexible this is. We now have running both Java and C++ at this point and they're finally communicating and they are running across boundaries and different Operating Systems. I think that they were running on the network and tests through Windows. So this is really a pervasive distributing parsing model. It's very, very novel and I believe that AmigaObjects will be really successful. We'll be talking about that in more detail later in the seminars."
Question 2: "Why is he using Linux as the underlying OS? What's being done to improve it? Will it lose it's Amiga-ness?"
Answer: Technical Difficulties
Question 3: "Isn't Linux big? How are we going to load all of that? What are we doing about that?"
Answer: Technical Difficulties
Question 4: Technical Difficulties
Answer: Technical Difficulties
Question 5: Technical Difficulties
Answer: "I think that this has been a particular issue that has been blown out of proportion, but I can give you an example. On a typical desktop system, for the types of tests that I've done, I can do about 100,000 process switches a second. That compares to about 1,500 on a standard Amiga. This is unbelievable fast hardware that is extraordinarily standard (400 Mhz). That's something."
"What gives you an idea what we can easily do, I think that it's important that we talk about real-time. For us, what real-time means: we don't really necessarily care about Nano-second precision, but we want our audio to play without any hick-ups, we want our video to play without missing any frames, and the fact that standard computers can run at 100,000 switches per second means gives you an idea that we'll have absolutely no problem playing real-time audio or video at much higher rates than what we see today."
Question 6: "What about networking? There's been threats about TCP/IP being lousy with Linux. What's really going on there?"
Answer: "I guess that the issue about TCP/IP performance has been probably more than anything, a matter of, I won't say ignorance, but the problem that most people have with Linux is that there's no single person you can ask about it than Linus [Turvold] and we can ask him more than no one else particularly can. It can often be difficult to really understand what the right answers are. I'd like to go down the list of these highlights.
Linux network performance in TCP/IP is actually among the fastest in the world. There's been some tests done run against WindowsNT and I think that you may have seen these on some Microsoft postings on the net, which has actually been rigged. Those of you know that Linux was actually mis-configured in there. So as a comparison, I think that the Linux system outruns the typical NT system when the Linux system was configured right, is somewhere between 2-to-1 and 5-to-1 in a typically test. That gives you an idea. Secondly, despite what I've seen on the net, I personally, and I have other people that agree with me, believe the TCP/IP implementation of Linux is one of the most stable in the world. The best argument for that is that about half of all the web servers on the planet is running Linux. If that was un-favorable we would have heard about it by now."
"I know that there are a lot of people having problems with the networking. Quite frankly all of the tests that I've done has shown nothing more than a world-class implementation. The speeds are incredible. I've already given you some of the network test speeds. It's ten times faster than anything that I've seen on Windows. So all of this stuff that you've heard about Linux TCP/IP about being inadequate, I believe is bogus, and all of our test indicate to the contrary."
Dr. LeFaivre added, "One thing that I'll mention is the stability. People have asked, does it make sense to have Linux in the home environment? Well gee, I can't think of a better place that you can run for months at a time and never crashing. We know of other Operating Systems that can't run for more than a few hours without crashing. When I was fine-tuning the slides in the room, a half hour ago, it crashed."
Question 7: "This is the X-windows questions. The native windowing system is X-windows. It's been around for a long time. People think that it's kind of a slug. What are we going to do about that?"
Answer: "Before I answer that, I also don't think that people will accept having to re-boot their TV sets. Just my thoughts. As a matter of fact we didn't talk about PARC tolerance when we talked about AmigaObjects but that's actually another element of the entire design."
"You're right X-windows is old. It's suddenly bigger than anyone would like. That is why we are only using it in some of our configurations. Something very specific. Only the appliance environments that Jim was talking about and what was on Rick's slides also, X is not the default windowing environment. X is what the enthusiasts and developers will be using for their development, which is very important to understand."
"On the smaller devices we have an Amiga windowing environment that's somewhere between what you are used to on your current Amiga, but updated for the types of issues that we need up at this point. I don't know how much you've thought about it, but for instance the way the old Amiga does video, the video doesn't sit on the bit plane. It's actually over-laid on the way out. So we need to see how video is being done in 3-D graphics. In 3-D graphics the video will come in through a 3-D graphics engine and will want to reside in the bit plane.
So the way that the windows manager, the way that you know it on the Amiga, like the layers and intuition if that's the way we want to call it on the Amiga, is actually being re-done and we're going to have a new Amiga windows manager that's designed specifically for the information appliances. It will have the same speed and benefits as the old Amiga did but aimed at this slightly different market. It's going to work great on the MCC as well, but my guess is, since the target for the new windowing environment is the appliances, some of you might, for development, actually prefer running Gnome, KDE, or what ever else you can find on Linux. I don't have any preference. It's really a matter of what we need to put on our information appliances."
Dr. LeFaivre adds, "And the other reason of course, to have it on the MCC which will have a desktop configuration is that you can run standard Linux applications coming from lots and lots of vendors. Which is kind of nice. Applications are good!"
Question 8: "What about standard Linux distribution? There's a lot of people distributing Linux. Is that all we're going to be doing is distributing Linux? What about applications?"
Answer: "I guess that both Rick and I have almost answered this question. We're not going to be another Linux. We are not going to be Red Hat with a Boing ball on it. Amiga/Linux, as a matter of fact, the AmigaOE is entirely separate thing. Linux will be a sub-set. Developers will probably get pretty full Linux distribution. Our appliances will get much, much smaller and quite frankly, people who will buy our appliance will probably not know that there is Linux running somewhere in it."
"I had a talk with a candidate that I was interviewing the other day. He was saying I know everyone is going to want to know what Operating System is running on their appliances. And I said Oh really do you really think that people are going to want to know what OS is running their appliances? he said absolutely. So I asked him, do you know what OS is running in your microwave? And he said that's a good question. I know that it's a silly example, but the truth is that most consumers don't want to know. They don't care. They want the thing to work. That's what we're doing."
Question 9: "Various questions have been raised about the hardware architecture, PCI, AGT, which graphics chip are we going to using, which sound sub-system, etc. and are there any other comments that you'd like to make? Although he's not really driving the hardware so don't blame him."
Answer: "Well let me just speculate. I can't say that I have any comments about Transmeta, and I can't say anything about that. I'll talk about graphics. I think that you saw from the technical brief that we teamed up with ATI. That's one of the best 3-D graphics chips on the market. I know that some people like 3-D effects but if you compare apples for apples. Most people don't really see the 3-D performance is absolutely terrible."
"The only one that remotely is close to ATI is GNT, the new one. So we understand that. I'd also like ask you don't get hung up on all the stuff about TCI and ATT. Because the thing about the Amiga architecture, you all know that there's much more to how a system performs than which bus is in it. So open your eyes and minds but don't get hung up on silly details at this point. We know what we're designing. We want a real-time multimedia system and what certain PC technologies systems that are out there right now doesn't have anything to do with the types of systems that we're going with."
Dr. LeFaivre adds, "Just one comment on behalf of all of the hardware engineers who aren't here, trust me, nobody is saying, Gee how can I design something that will be slow. They're trying to do their best to design something that will be cool."
Question 10: "OK, we're using all of this external technology, we're leveraging out the partners, that's really good. Where are focusing our innovation? I hope that that question has been answered tonight but maybe a couple of wrap-ups to finalize.
Answer: I guess, more the than any thing, I'd say that what the Amiga is going to bring to the table is the AmigaObject architecture. It's very important to understand that the AmigaObject is not one entity. It's the base on which everything is being built, so all of the capabilities of the AmigaObject will automatically be adopted by any other unit of execution and the system is extremely powerful."
"I'd say that the fact that we're adding streaming media, real-time 3-D graphics, and just generally real-time behavior to Linux is definitely an innovation at Amiga. A third one is that we're taking something Linux is really a workstation OS and actually making it work extremely well in smaller devices. It's an innovation at Amiga. And finally, and this is actually what most of our customers, and hopefully your customers are going to think most about, is the fact that we are creating a revolutionary user experience. Jim alluded to that a little bit but it's actually key to us that we get, I'd like to put it this way, I'd like to get the computer out of the experience."
"Most people want to use appliances, but users of appliances don't want to have to deal with the computer. That is the major contribution of the Amiga. We haven't really touched upon that yet, but we will at a later date."
Dr LeFaivre says, "OK, the architectures are fully defined. Most of the issues have been resolved. There are a few issues that we're still working on. That's the way that product development goes, there's always going to be a few things."
"One comment; we are building a great R&D team at Amiga. The people that are joining us, you seen so many announcements, as people have come on board, are outstanding people in their field; graphics, and object technology, etc. So there is great team building. We've got a hell of a lot of work to do. People are asking me about schedules and we've been out, but what we're shooting for is hard. People are working hard trying to get things on track. There's a lot of work to be done, and trust us, the team is working their asses off."
"We've got some great partners: people like Sun, Corel, Broadcom, some others we will be announcing at some point. Rumors have been circulating about others that we can't comment on."
"And finally, as Jim said, we are committed to being the leader in this next wave in computing. It's very exciting. It's an exciting time. A lot of people are rallying to this cause. Jim presented it in a much more inspirational way than I ever could but we're really very excited. It's great to be here. Thank you very much."