Two Interviews with Ken Kaplan, One of the Creators of OS-9

Two interview from two different time periods

Aug 24, 2024

The first interview I found was short, so I found a second. Enjoy!

from the October 1984 issue of The Rainbow

First Anniversary Special: An Interview With Ken Kaplan

Dale Puckett

Rainbow Contributing Editor

This month, our first anniversary column is a bit different. We have what we hope will be a special treat for you. While in Chicago for RAINBOWfest we attended Ken Kaplan’s seminar about the secrets of OS-9. Ken is the president of Microware Systems Corporation of Des Moines, Iowa, creators of OS-9. We tried to take good notes and you’ll find some of Ken’s most revealing comments and solid advice later in this column.

We also secured a personal interview with Ken. We had quite a few questions and we thought you would like to see the answers.

What’s the real story behind OS-9? Does it have a place in the future of the Color Computer community? What’s on the immediate horizon?

Here is the transcript of our interview.

Rainbow: Ed Juge told the crowd at the CoCo Community Breakfast this morning that if they wanted to maintain software compatibility on the Color Computer in the future, they should switch to the OS-9 operating system. Why did he say that?

Kaplan: I can’t speak for Radio Shack. However, there’s only so far they can go with the Color Computer hardware and still maintain 100 percent compatibility with Color basic software. OS-9 is different, because with OS-9 you don’t have to go around the system software to get the job done.

For example, there are a bunch of routines in Color BASIC that have been documented to do specific things — like get a character from the keyboard. But, there are also a bunch of undocumented routines. People who have written a lot of software for the Color Computer have used both the documented and undocumented routines. Some people have even gone out and worked directly with the hardware. That means that most of the applications software that runs on today’s Color Computer BASIC systems is pretty much locked into the present design. The programmers had to work directly with the hardware to get around the limitations in Color BASIC.

OS-9, because it has so many more functions, doesn’t force software authors to commit these sins. I think if developers use BASIC09 and OS-9 when they develop software in the future — and I think Radio Shack is saying the same thing — that Tandy will be able to come out with new, advanced machines that will run software developed on the older models. They will use OS-9 as a bridge to get from machine to machine.

Another thing that is happening in the market now is the realization that 64K bytes of memory is not enough. Memory is getting cheap now and 64K is just not enough — especially when you are doing a lot of graphics work. OS-9, unlike other software, Color BASIC and FLEX for example, can easily move beyond the 64K limitation. OS-9 can use 256K — even a megabyte — and a 19 or 20 megabyte hard disk. In fact, it thrives on it.

Rainbow: Radio Shack has announced a new licensing arrangement that will let software authors put OS-9 on a disk with a piece of application software. How does this affect the application software people who may now get inspired to do something for Color Computer owners, and how does it effect the user?

Kaplan: I don’t know the exact details. I know we have allowed Radio Shack to do this with our license. [Essentially], they are doing it for the end user. It should simplify things a lot. In the past, your customers had to go out and buy OS-9. Then, they had to set it up. When they bought your application, they had to copy all of your files to an OS-9 disk. Then, they found they had a problem with disk space and had to remove a few programs to make yours fit.

With this new license, application software houses can sell a completely turnkey program that consumers can plug in their disk drive and run right out of the box. They won’t even see OS-9. This will really simplify things for the user because the software house can pre-install everything.

Rainbow: Do you think Radio Shack’s new program will stir an interest in OS-9 and build a larger user base that application programmers can depend on if they decide to write software with OS-9 rather than Radio Shack Extended Color BASIC?

Kaplan: No question about it. It’s a pretty safe thing to do. In fact, the safest thing for any developer in the long run would be to do all development with OS-9. OS-9 on the Color Computer is growing by leaps and bounds.

Another thing that looks good for the future is that — without being too specific — there is a lot of action on OS-9 68000. There are a lot of companies moving toward OS-9 —including some well-known names. There's going to be a lot of 68000 activity and a lot of systems. And, there'll be a lot of applications software. Of course the high level languages like BASIC09 and C will run on both the 6809 or the 68000. That will broaden things even more. So, if you write an application for the Color Computer now, you can move it to the 68000 very easily.

Rainbow: In other words, it would run on any of these new machines — those “big names” that will help expand the OS-9 user base. Will these new machines be affordable?

Kaplan: The prices will vary. Some will be very affordable — home systems. From there they will run all the way up to the high ticket professional systems.

Rainbow: Almost everyone is talking about the Sinclair. Is OS-9 68000 headed for that computer?

Kaplan: I can’t say specifically at this time. I just can’t announce who our customers are.

Rainbow: What about the Macintosh?

Kaplan: We’ve got a Macintosh and we’ve played with it. It's a very nice system. I think Apple’s philosophy is to keep the Macintosh locked very tightly in their control. They want to control all the system software and all the hardware. That’s why they designed it so that it’s not really expandable. You can’t plug in extra cards like you can on the Apple IIe or the IBM. I think they are encouraging people to write applications programs for it but they’re not encouraging companies like Microware.

Rainbow: I understand they want all application programs to work through their windows. What would be the advantage of having an OS-9 type operating system in that environment?

Kaplan: I don’t know. The present Macintosh software is not multitasking. It does not have the UNIX type features. It has sort of a tree structure file system but I find it sort of confusing, the way the file system works. It could be cleaned up a little.

Rainbow: One of the things that has been a mainstay on the Color Computer has been the graphics capability of Microsoft BASIC working with the SAM and VDG chips. Are there any plans for the future that would expand that graphic capability to OS-9 or BASIC09?

Kaplan: Yes, there’s some discussion about that right now. I think you’ll see future releases of OS-9 that will be more capable. Some of the things that are being kicked around right now include a paint command, a draw command, and a built-in Hi-Res graphics package.

Rainbow: Would that run at the assembly level?

Kaplan: Yes, but it would be accessible from BASIC09.

Rainbow: Your GFX package is I-code, right?

Kaplan: Yes, but you see GFX actually just drives the operating system. A module within OS-9 contains the graphics routines. They are activated by control characters. This means that you can even display graphics from the Shell. GFX just gives you a cleaner interface. You don’t need to know all those control characters. GFX is a little translation module.

Rainbow: What do you think is the OS-9 user base on the Color Computer?

Kaplan: I’d say approximately 50,000, maybe more.

Rainbow: What’s the problem with OS-9? Look at the magazines. RAINBOW is the only one that covers OS-9. They say there are not enough users because out of the million CoCos sold, there are only 100,000 disk drives. That limits the usability. What’s the answer in the long run? Will OS-9 ever get to the consumer level — I mean like Commodore?

Kaplan: I think it’s an illusion. First of all, I believe that far more than 10 percent of the Color Computer owners use disk drives. I would guess it’s more like 40 or 50 percent. I don’t buy that 10 percent figure.

Rainbow: Thank you, Ken!

Kaplan: Thank you for all your support in THE RAINBOW.

Ken Kaplan Addresses RAINBOWfest Crowd

Just before the interview above, Ken Kaplan presented a few “OS-9 Secrets” to several hundred RAINBOWfest attendees. He opened with his perception of the mistake some people made when they bought OS-9.

Kaplan: People go to their local Radio Shack store in the mall, buy OS-9 and take it home. They take it out of the box, plug it in and then say, “What’s it going to do?”

Well, OS-9 is wonderful if you want to try to write an assembly language program. But, not that many people have that in mind. And, it is a little complicated.

I think most people like to do, want to do, the kinds of things you can do in Color BASIC. Except they want more features, more capability. So really, you need to think about OS-9 not as an end — a primary tool — but rather, you must think of it as a foundation.

OS-9 is the basis for running other programs. It gives you a very powerful foundation. All features in OS-9 allow people writing applications, or yourself when you write a program, to use advanced programming concepts and techniques to get the job done more effectively. And in fact, in some cases do things you couldn’t do otherwise.

BASIC09 Is A Special Tool

BASIC09 is also of special interest to you. It’s another layer that goes on top of OS-9 and will probably be the main day-in day-out tool you will be using.

BASIC09 is a very, very powerful BASIC. And, we’ve added some things to standard BASIC09 to make it even more useful on the Color Computer. Graphic commands, joystick inputs and things like that are a perfect example.

The neat thing about BASIC09 — besides the language itself — is the fact that you can get to any operating system command from inside BASIC. When you’re in the command mode, you can type a $ and run any OS-9 command, just as you would if you had booted the system up and never gone into BASIC.

Also, you can run any of the operating system commands from statements in your program — and that’s a very powerful capability. In fact, a lot of people who have used OS-9 have the start-up file automatically load and execute BASIC09. They do all of their work right inside BASIC09 and never need to talk to OS-9. That’s a very good way to do things.

About Those Crowded Disks

One thing that we’ve heard a lot of comments about from people who have had some difficulty getting BASIC09 going, has to do with the fact that there’s so much software crammed on that system disk you receive from Radio Shack. In fact, there’s so much software on it, that BASIC09 won’t fit. As a result, a lot of people have tried to put their BASIC09 disk in drive one and leave the OS-9 disk in drive zero. When they do this they run into several problems.

I recommend that you make two backup copies of your system disk, and then put the one you bought away for when you crash the disk, or your cat eats it, or your child spills a bottle of milk on it, etc. Next, label one of those backup disks as your BASIC disk and label the other one as your assembly language disk. If you get around to doing some assembly language programming, your assembly language disk will be there ready to go. It’s got everything you need on it when you take it out of the box.

If you’re going to work mainly with BASIC, you need to start by deleting some of the files that are on your disk. I would delete the commands that are designed mostly for machine language programming. That will free up a lot of space. In fact if you want, you can delete the editor and assembler program. After you do this, BASIC09 and the two other files — the GFX module and the INKEY module — will fit nicely. You’ll be ready to go.

As you move further along, you’ll notice a few other sort of arcane utilities on your system disk that you won’t use or very rarely use. You can get rid of these, too. As long as you’ve made a backup disk, you can feel free to clean things up.

Once you realize that OS-9 is a foundation for building software, you realize that the programming languages available on OS-9 are pretty important. Aside from BASIC09 and C, which are the two main languages available for the Color Computer right now, there are several additional Microware languages coming soon.

Later this year you will probably see a PASCAL compiler. A FORTRAN compiler should follow sometime toward the end of the year. Later, a COBOL compiler will be available. These are all very good languages and should cover almost any kind of program that you need to write.

OS-9 Is Important Because It’s So Much Like UNIX

The main thing about OS-9 for the average person — and the reason it’s a real step forward — is that it is closely modeled after UNIX. I think the world is beginning to realize that UNIX and UNlX-style operating systems are the future.

The Color Computer is a very extraordinary machine. It is the only machine in the world I know of that will give you the entire UNIX environment for under $1,000 — including a full-blown C compiler. This can give you a tremendous advantage, particularly if you’re a student. It gives you a little taste of what tomorrow is going to be like.

A Myth

A number of myths have been promulgated by the press and others about OS-9 because it’s so new. One has to do with the issue of Level I vs. Level II. Level I came out about a year and a half before Level II, right? I used Level I for a long time and the fact that there’s a memory restriction in Level I that keeps it from being useful is absolute fallacy.

Most people who got started on OS-9, used Level I for a long time. There’s plenty of memory on the Color Computer to do all sorts of things with Level I. In fact, there are methods in BASIC09 that you can use to make better use of available memory because things are modularized. You can chain programs together and so on. That’s much easier and convenient to do in BASIC09 than in any other software you can run on the Color Computer.

I like to use the analogy of buying a car. You don’t go out and buy a Greyhound bus so you can take all your friends when you go somewhere. Rather, you buy a car that suits your needs and has the right capacity for you right now.

There’s no reason why Color Computer OS-9 can’t do that. I use it a lot at home, all the time, and I very seldom find myself getting frustrated by memory restrictions.

Another Myth

Another common myth about OS-9 has to do with its multitasking abilities. It’s sort of extraordinary that a low cost machine like the Color Computer can do multitasking.

Then, people say, “Well, why would I want to plug three or four terminals into a small computer?” or “How could I? Is it really practical?”

The answer is, for most people, no. You can. practically speaking, run a couple of users on a Color Computer. This is timesharing.

Multitasking is not the same thing. There are a number of applications where multitasking on a single user system could come in very handy. Here’s an example.

Suppose you want to set up a home security system with your Color Computer. You have sensors around the house and you know how to write the program in BASIC. You turn on everything and run your program. It works.

Then, several hours later you decide you want to use your computer to play some games or need to use your word processor to write a letter. At this point you discover that if you do this you have to turn off your home security program. Y our computer can only do one thing at a time.

OS-9 frees you of this restriction. It lets you run your home security program as a background task under OS-9 while you use your Color Computer to do other things at the same time. This type of application is very valuable, even on a small system like the Color Computer.

OS-9: Twenty Years and Going Strong

from the August 1, 2000 issue of SD Times

BY ALAN ZEICHICK

The Midwestern roots run deep at embedded systems veteran Microware Systems Corp. The Des Moines, Iowa-based company, best known for its OS-9 real-time operating system, has been hammered on Wall Street, recently trading more than 75 percent off of its 52-week high. Yet Ken Kaplan, Microware's co-founder, chairman and CEO, is both realistic and optimistic about the company's (www.microware.com) future.

SD TIMES: What's happening with Microware's financial status, and what does Motorola, which owns 11 percent of the company, think about the current situation?

Ken Kaplan: Motorola invested in us about five years ago, and they're also a major customer. We provide software for a number of their products, such as their two-way pagers and interactive set-top boxes. I've never heard anyone at Motorola complain about the stock price — I don't think they care that much.

In the larger picture, Microware has always been hanging over the edge of the future, and sometimes a little bit too far. We've been ahead of our time. When we introduced OS-9 back in 1980 to 1981, it was 10 years ahead of its time back then. We're still an innovative company, but in the past few years, we decided to invest in being best-in-class in particular vertical application areas. We've invested heavily in product and business development.

At the same time, we went through the same trials and tribulations that many small companies go through in the post-IPO phase, and the stock oscillated almost in a sine wave. When the Nasdaq went through its extreme craziness in March and April, we peaked at over 11 time, says Kaplan.

[Editor's note: a high of 11 1/2, closing at 10 9/16, on March 7], and that was an unrealistic valuation. I can't say our stock shouldn't have been there, but given our underlying position, that wasn't a supportable stock price.

Do you see the company's overall financials being where you want them to be?

Our objective right now is to get the company into the black. We're trying to do that as quickly as we can. We did a financial deal a couple of months ago [Editors note: sale of $3.5 million in preferred stock, with the option to sell $3 million more, completed on April 19], and the reason we did that was to send a message to the market. Some of our competitors had been spreading some gossip and innuendo about us and our being public, and we wanted to send a message that one of the pluses of being a public company is having access to the capital markets. That was an important confidence builder and should take care of our financial needs for the immediate term.

Internally, the company is doing the right things. For the past year, we've been getting our message out. We have invested in not only product development, but also marketing to get the story out about us, our products and our capabilities.

You mentioned an emphasis on vertical markets. Does that mean DAVID [Digital Audio/Video Interactive Decoder] and the HDTV products?

That's one of the confusing things. We've been very well known for our digital TV. We still have a lot of business in that market. But what Microware is about is Internet appliances and connectivity. A lot of people, rightly or wrongly, think about operating systems as commodities. They're not, and OS-9 has some significant advantages that our competitors don't have. It's an important foundation for us to build complete solutions, particularly in embedded computing and communications convergence; that's our great strength.

Will Microware be putting more of an emphasis on vertical solutions, or will you be maintaining OS-9 as a general-purpose RTOS?

We're not going to turn away any business. But many of the higher visibility things we do are going to be the complete solutions that we build. It's surprising, but there's a lot of commonality today. Just about anything that has embedded intelligence needs to be connected to the Internet. The key strengths we have in communications software, in Java, in HTML — those technologies are increasingly common in even what may be considered mundane industrial applications.

Should I think of Microware as a product company that sells development tools and RTOSes, or as a service company that builds complete solutions?

We have a growing service business, which provides architecture and design through specific engineering services for our customers. But when you talk about a "solution," it doesn't necessarily mean it's a one-off customized thing. We have a product line that's like tinkertoys, like building blocks. Rather than have to go through the expense and time to customize a product or application, we can help our customers create a very significant percentage of what they need through using off-the-shelf building blocks — often as high as 80 or 90 percent. And then the customer, maybe with our help or a third party, can build that last 10 or 20 percent which is very specific to their product. That's a lot different than with competitors' products, which is usually much less than 50 percent off the shelf. It's all pre-integrated, street-tested, ready to go.

Are there a lot of third-party tool vendors selling integrated development environments, libraries or tools for OS-9, or is it that if you're using OS-9...?

The tools aren't an issue right now. Two years ago, some of the companies in our market decided that tools ought to be a battleground. There was a lot of talk about "open tools" and "open IDEs," and there was a lot of baloney floating around. Open IDEs? None of them are open. Now, things have gone full-circle, and a lot of people are talking about the benefits of OS-aware tool sets.

When you have a good, solid tool set that's designed for a particular operating system that understands how the operating system operates, and lets designers see inside the OS, as long as it provides all the basic functions, that's what the customer wants. We have a solid IDE called Hawk, which we've been evolving for many years. It is built on a framework from Premia called CodeWright, a very popular editing and source-code generation product.

Also, the more we go toward Java-based and HTML-based applications, products can be developed on PCs under Windows and the IDE becomes less and less relevant anyway.

What about Java? OS-9 currently supports PersonalJava, but doesn’t yet support the full Java 2 Micro Edition.

We will be rolling out our KVM implementation by the end of the summer.

J2ME is pretty new. We've been staying at the bleeding edge of Java; we're in a leadership position. Despite all the hype, everyone's paced by how fast the ink on the specs dries and how long it takes Sun to get their reference implementations and code packages out. We have shipped more real embedded Java systems than everyone else combined and we intend to maintain that leadership advantage.

Do you see Java being the predominant way of developing for OS-9 and other embedded systems in its class?

For those applications that will be distributed over networks and need to be system-independent, Java is a very critical technology, and it's good for us. I don’t think it will, in the near term, handle hard real-time, though one of our engi of the developers of the Real-Time Java specification, which was just finalized and published When that happens, Java will be able to handle the entire gamut of applications, though there will be some issues of efficiency and performance because Java is largely an interpreted language.

C is going to be around for the foreseeable future, but Java development will at some point get to where it's on a par with C development.

It's interesting. Take companies that are single-trick ponies-let's say all they do is Java. They are at a disadvantage. They may run on 18 different operating systems but don't necessarily integrate well with middleware and other stuff.

Is there someone in particular you're referring to?

Yes, Sun. I think they would admit that they don't expect people to take their off-the-shelf JVM and really use it; that's why they need JavaOne. I don't think they expect that the code they distribute is production code.

There's a lot of talk about embedded Linux. How does embedded Linux compare with what OS-9 is today? What will your market be like when embedded Linux does everything that OS-9 can do?

There's a strong sense of déjà vu here, Alan. When we originally wrote OS-9, we studied other operating systems, and we had heard about this operating system called Unix. We decided that it was too big, too clumsy and too inflexible for industrial embedded application, plus it wasn't real-time. So we built an operating system that had a lot of features like it and the same basic architecture, but was designed to be modular, embeddable and real-time.

The bottom line is that Linux is a modern incarnation of Unix-it's a clean-room Unix. It has the advantages and disadvantages of Linux but also has a common architectural heritage with OS-9. We've had customers take C applications from Linux and recompile them to run on OS-9 and vice versa. You can't do the real-time stuff on Linux, but we can peacefully co-exist with Linux. We will see customers using Linux for the main application engine, and OS-9 as a real-time front end for it, and maybe connect them over a network. We'll see some competition on the high end, but fundamentally it'll be hard to make Linux handle hard real-time and get into a small memory foot-print. Getting people to agree on how to do that will be harder still.

Will Microware offer tools for embedded Linux?

Probably the next step we'll do, though we've not formally announced this, is make our IDE available on Linux-based development stations. If we see a market, we could get a lot closer to Linux and be highly compatible with it. We'll keep an eye on it.

Looking forward five years down the road, what will Microware look like? A much bigger company, say?

We don't expect to be another Microsoft. We think that building software requires craftsmanship, and there's an optimum size for a software company. Our objective isn't to be huge we might grow horizontally, or find companies to join forces with that have complementary products. We're going to concentrate on communications, embedded systems and Java, building superior products and superior services. I don't think size equals quality.

Wind River has been on this huge acquisition process. I don't hear many customers think that that's necessarily a lot of benefit for them. They're very large, very diverse and an aggregation of all these smaller companies.

What about being acquired by Motorola, or are you better off as an independent?

I don't think it would be good for a chip manufacturer to acquire us. The reason is that would make us partisan. We support most of the popular microprocessors: Motorola, Intel, MIPS, Hitachi and others. The minute one of those guys owned us, we'd probably lose our ability to effectively support and market into those others. A partnership or alliance with a company that would be more compatible, like another software or tools company, that might make sense. The real value's not to the chip manufacturer.

Computer Ads from the Past

Discussion about this post