Ever wonder what the world of computing would be like if the IBM PC was originally created in an atmosphere of open-source?
The IBM PC, the personal computer that launched an industry of over $500 billion in combined stock values, will be 25 years old this Saturday (original IBM press release.) It is estimated that there are about 1 billion PCs currently used around the world today. With the unveiling of the IBM 5150 (well... they had to give it a number!), IBM made a statement that personal computers were important; so important that IBM must throw its hat into the consumer ring. Of course, IBM was not the only company making personal computers at the time. But, they were the largest.
Much has been written, and will be written, about the technical success and failure of the IBM PC. But, there is no argument that it was, and is, a massive economic success. The un-IBM-way that the PC was created had much to do with its success. Unimpeded by IBM bureaucracy, the Boca Raton team created a computer with off-the-shelf parts. IBM's decision to publish the architecture was a brainstorm that resulted in a flurry of software upstarts and later new hardware add-ons that propelled the PC to stardom. It is well known that this decision was also their own economic seppuku.
My question for today is: what if the IBM engineers worked in an open-source environment? What if they took advantage of other work available to them? Well, in one way they did. They used non-IBM components (e.g. the Intel 8088) and went outside for rudimentary software (Microsoft for operating system software, BASIC, Assembly, etc.) That's a bit advanced for the time... especially for IBM!
But, what if they went further and tapped technology centers such as Xerox PARC (who had ethernet, laser printers and bit-mapped displays running in 1976) or the work at SRI? At the time, Apple was creating the Lisa/Macintosh that utilized PARC innovations. Reports have indicated that IBM had never visited PARC. Maybe they were never invited. (see Andy Hertzfeld's perspective from of the Mac team, as he recalls the IBM PCs launch.)
Not ignoring the business hurdles, I'd love to hear the technical insights of O'Reilly readers. How would those 1 billion PCs be different if IBM utilized open-source avenues to create the original? Would the growth of the personal computer market be changed dramatically, or insignificantly? Would Apple be as successful as it is today (sans iPod)? What's your take on it?
See more about personal multimedia at my site: brainFlakes
Technorati tags:
technology, open source
Categories
Web
Read More Entries by Brad Fuller.
Become an O'Reilly Member
Visit the Digital Media Forums
Subscribe to the Digital Media Blog Feed
Subscribe to the Digital Media Newsletter
Did IBM look at PARC? I don't know, but I believe Steve Jobs did tour PARC and got important ideas, e.g. the mouse and GUI. IBM of course studied Apple before entering the PC market, so the influence could have been from that direction.
Do we know that they didn't look at PARC? It would be interesting to know if they knew about the work being done there but for some reason decided to intentionally go in a different direction.
I suspect one reason might be economic. They seemed to want to be the "Model T" of computers, as did the original Apple (before Lisa/Mac).
I think to most users who buy PCs and software, the notion of open source is a non-entity. They will most likely buy a brand name because it's something they've heard of and/or know others who use it.
Is a definition of "revolutionary" in order, or at least what constitutes the threshold from evolutionary to revolutionary? I suggest that max/pd is a bit revolutionary and that came originally from one guy (although I admit I don't know the involvement of the IRCAM crew.) Engelbart has a patent on the mouse and is the only one named on the patent. Check out NerdTV's interview with Engelbart -- I remember him saying how hard it was to obtain resources!
I think you would agree that invention is an individual affair. An inventor is often alone, toiling to make something work. Innovation involves the interaction with others and economic incentives. Both innovation and invention are required for technological progress. But in this particular case, the question isn't about innovation or invention; evolution or revolution. The question is, what would the IBM PC be, if company walls were knocked down between creators?
One might reasonably invert the underlying premise of this thread. Is it possible the an atmosphere of open-source might have actually crippled the development of the PC platform ?
Fundamental research, as opposed to re-engineering, takes money. This money was generally supplied by academia and industry, and was essential to creating an environment so talented researchers had the time to invent the mouse, ethernet, laser-printing, and the modern desktop GUI.
The open-source community has, for whatever reason, produced a string of extremely high-quality, but ultimately non-revolutionary products. Do we really think this distributed development model of talented volunteers would have created the mouse ? Or was the mouse the result of a single researcher having a flash of innovation ? Granted, we are now blessed with a variety of workable GUI desktop alternatives, a great Office clone, and an OS which has been ported to every hw platform in existence. This is substantial, but in the end, all of this seems more like incremental re-engineering of existing work - as opposed to revolutionary innovation.
To create ethernet, you need lots of free-time. As the cliche goes, time is money.
Sincerely,
Craig Hansen-Sturm
For another view, you might pop over to Tim O'Reilly's blog "Open Source: Architecture or Goodwill?" on how Web 2.0 eliminates the challenges of software distribution - be it open-source or binary. However, don't forget to read the comments for legitimate challenges to Tim's viewpoint. If you track backward, one could imagine how different the original IBM PC development might have been in a Web 2.0 world.
The PC arrived in a world that had three main open source communities:
All three were facing pressure to close up as a way of making money (in the BASIC case that normally meant the programmer moving on to assembly) and when this killed 1 Richard Stallman decided to act to save 2 while bringing with him at least some stuff from 1 (only Emacs and the gcc hardware description notation actually made it).
To enter the microcomputer market IBM did what is now called "benchmarking" (imitating whatever the leaders are doing), so the cross between an Apple II and a CP/M+S100 machine was just about perfect. The BASIC in ROM and the option of a cassette interface represented an effort to tap into 3. But the PC was adopted in businesses and not so much in homes which caused this to fail. By the time clones started to make an impact they didn't even bother copying these features.
Smalltalk had already run on 8086 machines back in 1978, but these were considerably larger configuration than early PCs. Digitalk's Methods and Smalltalk V never ran on anything with less than 512KB, for example, and the original PC couldn't even be expanded to that much (the XT could). And don't forget the performance issues - the PC clones killed the DOS computers because applications that used the BIOS to draw to the screen (like Visicalc) were considered too slow compared to those that accessed the hardware directly (like Lotus 123).
Note that I was seriously thinking about all these issues back in 1982 and that gives me a slightly different perspective from someone just looking back today.
What I really wanted to explore was the software angle. Putting aside hw cost for a moment, let's say Squeak was available at that time in open source form. It runs on low powered computers and is based on Smalltalk-80. Would IBM choose this over Microsoft? If not, maybe Squeak would become popular with developers because of its usefulness? - in a grassroots sort of way. Note that the benefits of OOP and Smalltalk were known in the software community to some. Curiously, "Design Principles Behind Smalltalk" was published in Byte in August 1981 - the same month the IBM PC was announced. If Smalltalk was available on the PC in 1981, or even on the MAC when it came out, software development most likely would be very different today. And, as a consequence, users would use the PC much differently.
The earliest IBM computers were open source in the sense that the IBM gave the source code for the operating system away with the computer. IBM kept the rights to it, so it wasn't free, and in fact the US government forced IBM to sell the software separately from the hardware once other companies started to sell "IBM compatible" mainframes.
The reason that the PC used an operating system written by some other company was because the PC was made by a renegade part of IBM and the people who made big computers at IBM had no idea what was going on and so couldn't put a stop to it. The PC was not part of IBMs strategy at the time. For example, about the time the PC came out, there was another part of IBM that sold computers that ran Unix. I think they used the Motorola 68000 as their processor and I remember that they were sold to scientists to be used in labs. They were much more expensive than the PC, partly because they had a disk and more memory and so needed to be, and partly because they were being sold as "computers" and not as consumer electronics. But the original PC came with 64K of memory and no peripherals other than an audio tape player that you would provide yourself to store your programs. It was just barely a computer.
PARC started selling the Xerox Star about the same time that IBM started selling the PC. They charged on the order of $20K for it. That was more than the salary of a secretary at the time, so it didn't sell very well. It was closed source, too, and it was hard to add components to it. I used one in the early 80s, though, and they were beautiful and easy to use, a marvel! If they had sold them for $6000 or $7000 then they would not have made any money on them, but they could have built up the market and set the standards. Xerox would now be Microsoft, IBM, Dell, Sun and Apple all rolled into one. IBM was no more prescient than Xerox. They were just lucky that they were big enough and disorganized enough that some renegade group could introduce a product than any of the top IBM managers would have instantly killed if they had known about it.
Ralph Johnson
All those IBM PC clones, a few years later, were the perfect eco-system for open source. Would have open source spread if there weren't "a computer on each desktop" (mis-quoting Mr. Gates).
There were BSD efforts for this 'PC' platform soon enough, but the real success came from Linus Thorvalds, porting Unix to it.
Ever wonder what would be the atmosphere of open-source without IBM and Microsoft bringing computer to the masses ?
Best regards,
John Mettraux
http://www.openwfe.org
Yes, IBM was open source to a great extent, more than previous pc's. They were not like Tandy, Texas Instruments, Atari, Apple, and others who tried to control the entire system. IBM knew the beauty of standardized components. Just try swapping a Tandy hard drive with an Apple IIe sometime and you will suddenly see that IBM was on the right track. Of course they didn't see everything. For example, that the nerdy guy selling the operating system, Bill whatshisname, was going to make a killing by retaining rights to his software. IBM could have bought him out then and there.
Remember how the old PCs used to have an origin date of something like Jan 1, 1981? I've always thought a science fiction writer could use that as a new dividing point for time: BPC and APC. Before PC, After PC.
Peter Tocco
www.plainview3d.com