Slashdot Mirror
IBM's Purple Book and Open Source
Bill Kendrick writes: "I noticed a ZDNet article titled "Why we should hail IBM's ode to open source--the Purple Book". It compares IBM's open release of the classic PC's hardware and BIOS specifications with today's OpenSource model and Linux." Shortly after IBM's open-spec PC, they reverted to the closed PS/2 with a patented bus in an attempt to monopolize the exploding market. Hopefully this particular bit of history won't replay itself.
These guys have taken the ball and run with it, as far as Linux is concerned, but let's not deify the group that brought us Microchannel architecture in a move to regain absolute control of the market. Working with IBM in Linux development is good and important, but don't lose sight of their history as a megacorp bound on dominating everything in sight. The only reason they aren't still doing it, is a bigger, meaner and more evil company came along.
I gotta get a tight tension on...
Hey, They _were_ an evil company. But they found religion. Maybe people (and corporations) can change... Besides, they threw a great party at NYC Linuxworld...
Acquiescence leads to obliteration
Originally IBM made equipment to deal with punched cards. However setting up a tab machine was very time consuming. Early IBM business computers were basically automated plug boards, they still used cards as i/o, but the program was quickly and easily changed. It wasn't until the 70s that mass storage started to replace punch cards. Because of this, mainframes use EBCDIC which is an enhanced version of the original punch card code, almost totally backwards compatable.
Punch cards never used ASCII, first they were in use since about 1880, long before ASCII was thought up, secondly, ASCII isn't suitable for a punch card code - if you tried to punch a card full of '7' characters, you'd end up with 400 holes on one card, which wouldn't have any structural strength. Punch cards had numbers encoded as a single hole, and everything else as one or two holes, giving a maximum of 160 holes possible on a single card. This gave a maximum of 64 different codes, so when the computer read in the card it could be very easily translated into a six bit code, a Binary Coded Decimal Interchange Code or BCDIC. Extending this to 8 bits gave EBCDIC. Here is a good description of card formats and EBCDIC.
You're absolutely right. IBM made a mistake, and now they're back to being the fuzzy little happy company they used to be back when they were founded by a kind old man with a heart of gold and his loyal wife out in America's heartland.
We're lucky that IBM has returned to its pure intentions by spreading Linus' Good News to the world, and doing so completely selflessly. As the prophet Stallman says,
In conclusion, I think we can all rest easy in the knowledge that we all have a multinational ally in our blessed Jihad against the anti-christ and his unbelievers at Microsoft. They are an evil, self-interested demon-corporation.
IBMah Akbar!
Hopefully, this little bit of history will replay itself.
.NET framework. They're putting the squeeze on, in an attempt to lock it all up once and for all. Hopefully the market will respond to .NET in the same way it responded to MCA: evil empire loyalists may adopt it, but everyone else will go running to some other, non-empire-controlled, open standard(s). J2EE, Linux, it's all there and it all works.
IBM took an open standard (the ISA bus) and tried to take it closed (the MCA bus) so they'd put an impenetrable lock on the market. The industry responded by saying "to hell with you, IBM; we're all going to EISA (and later, PCI)." Who's playing the close-it-up game now? Microsoft, of course. The Internet currnently runs on open standards, but Microsoft's 'MCA bus' is the
Tired of FB/Google censorship? Visit UNCENSORED!
If I remember my history correctly, the original IBM PC was open-spec only because they didn't have enough time to come up with something proprietary. They wanted to monopolize the market from the start, but they were running behind and had to get something out so as not to lose the market entirely. So, I don't think we have to worry too much about this piece of history repeating itself because their push for openness isn't motivated by time pressures this time (at least I don't think it is).
-----
Free P2P Backup, Windows & Linux
THE PURPLE BOOK "contained the hardware schematics for the IBM PC as well as the code listings for the ROM BIOS," Dave Bradley, one of the machine's 12 original designers, later explained to me. "It contained just about everything you'd want to know if you were going to build a device that would plug into the IBM PC."
In the Purple Book, as Bradley said during the panel, "We told all the PC secrets."
IBM wasn't the first personal computer maker to spill its guts. Apple published the source code for its Apple II. Atari and Commodore also offered similarly extensive documentations. But for Big Blue, a company that built a dynasty on proprietary products, the Purple Book represented a break with tradition as almost as radical as Martin Luther's breach with the Holy Mother Church.
WHY DID IBM SO WILLINGLY bare the soul of its new machine? Bradley again: IBM wanted to "make it as simple as possible to design hardware and software that would work with the PC."
"We wanted the software and hardware industry to participate."
Participate they did. What's more, the Purple Book made the IBM PC easy to copy, and thus, in came the clones. The result: A de facto standard was born, and that standard made way for the widespread deployment and use of PCs. The rest, as they say, is history.
The historical significance is the parallel that exists between the Purple Book of yesterday and the open-source movement of today. The comparison isn't a perfect one. The Purple Book did not constitute a license for use; IBM retained intellectual property rights.
Whatever! The retaining of intellectual property rights is ther whole point. What they did is what everyone else who had attempted to put out a PC would have to do in that era. The subset of technicians working on these technologies was quite small- small enough that a collegial flow of information was necessary even to drum upo interest in one's hardware.
So what IBM was doing was trying to raise itself to a playing field which Apple and Commodore had already delineated; to break into a technological community which was already occupied with other hardwares, it had to disseminate technical information.
There is a parallel today; Geron, the company which licensed the technology to extract stem cells from blastocyst-stage embryos, dissseminated the technology, advice and support to institutions of learning, retains commercial rights to any salable products that come out of these laboratories - or even the precursors of those products.
Then and now, such a technique is to take advantage of an academic desire for learning, or a desire to help the sick, and commercialize its output.
There is really no choice for software developers in the Microsoft world, or for stem cell scientists outside the apron of federal approval, except to sell their first-born breakthroughs to loan sharks.
I've said it before and I'll say it again; capitalist systems cannot sustain innovative energy or scientific responsibility.
Goat sex free since 2001
Simply put. Microchannel was great. The biggest problem was the stupid reference diskettes, and that was simply because flash memory wasn't there, yet.
Microchannel machines simply worked longer and more reliably than ISA or even PCI machines.
The problems was the stupid #$%& licensing terms. Gotta separate the technical side from the idiot marketing side.
The living have better things to do than to continue hating the dead.
and licensing their technology. IBM has more patents every year than any other company. If I am not mistaken, the PS2 architecture was ahead of it's time by having devices speak directly to the processor and memory speeding things up. I also believe IBM opened the PC spec up to put down the monopoly abuse cries from their trials. If memory serves, IBM put alot of companies in business.
Yes, I realize IBM has said that they won't do their own distro... but I still think they should.
One of the biggest reasons for the success of the PC was not just the openess, but because the IBM brand name was something that provided a bit more confidence for "the corporate suits" to adopt... an image which Apple and Radio Shack didn't quite have then... and most distros don't quite have right now.
IBM should probably buy an existing distro like Redhat in addition to the multi-distro support they're already doing. Thanks to the GPL, all of their developments for their own distro will still be available to other distros who will in turn refine IBM's developments, and so on.
Publishing the interface specification has nothing to do with creating an "open source" computer. The Purple Book provided enough information for third-party developers to build adapter cards compatible with the PC, but not enough information to clone the PC itself.
IBM was simply imitating the highly sucessful Apple II, which Wozniak designed as a platform that his fellow hardware hackers could easily extend. IBM also imitated Apple in keeping the ROM BIOS source code closed, and making it legally difficult for anybody to reverse-engineer it. Neither company was interested in having competition build the basic platform.
Unfortunately for IBM (and fortunate for consumers of commodity hardware), Phoenix, Compaq, and others were able to use "clean-room" techniques to reverse-engineer IBM's software without breaking existing law. Alas, the law has since been tightened.
Did ZDNet lay off it's editors?
The author needs a history lesson. I won't even bother to correct his "Bill Gates wrote DOS" misinformation.
But, the part that really needs clarifying is: the "Purple Book" ignited the PC revolution. While it is true that Microsoft owes it's monopoly to an open hardware platform, the "Purple Book" was not the key.
IBM owned the motherboard and system by holding the IP to the bios. People could build on top of the PC platform with the "Purple Book", but only IBM could build the systems -- Until Compaq (and soon after, others) reverse engineered the bios.
Then anyone could build motherboards and systems, and the PC revolution was ignited; hardware became more functional and faster at a furious pace.
If IBM still controlled the platform, then 8088's would still be hot technology.
And here is where the real lesson for Open Source starts.
If today's IP laws favoring those with popular software were in place in the 80's, then Compaq would never have been legally able to reverse engineer the IBM PC bios, and IBM would still be in control of their platform.
Open source indeed has the potential to put all software application's competition on a level playing field (where no one vendor can leverage the operating system to favor their applications and break other's applications).
But, in order to get people to switch from Microsoft to Open Source, we need some degree of compatibility. Customers are slow to change old habits. They fear training. They fear that old data won't make the transition exactly correctly. They trust that Microsoft won't shoot itself in the foot as often as they kill their competition... usually Microsoft applications work with their OS, as long as you keep upgrading on their schedule. If you want to gauge how quickly the US market will switch to an Open Source OS, just look athow quickly the US has embraced the metric system: it may be a world standard,it may be superior, it may be compatible, but the thought of change scares people.
Compatibility with Microsoft requires reverse engineering of their API's, file formats, and protocols.
Competitors are used to chasing Microsoft's tail: every release and patch has changes that make compatible software obsolete, and competitors have to scurry to be compatible again. But, this game just took a turn forthe worse: with today's IP laws, Microsoft can patent a portion of their protocols, API's, or file formats, and make compatible competition illegal. They don't need to patent something novel, just different: Microsoft customers are forced to follow like lemmings.
From the recent ".net" news, it looks like that's exactly what Microsoft intends to do with Samba. Samba is usually a foot-in-the-door for the Linux OS in corporate america. It's easy to show management the savings on a Linux based file server running Samba.
Now, Microsoft is going to require a license for the encryption algorithm for their password verification and modification.
That will kill Samba.
No Open Source project can afford a license. Usually when pressed fora license, the Open Source project ends compatibility immediately. For Unisys's LZH patent on GIFs, we just switched to other image compression algorithms. We've still yet to see how far Dolby will go to stop distribution of Open Source AC3 decoders. The mere letter from the lawyer is usually enough to stop anOpen Source project.
Worse: even if Microsoft were to grant a license, it would probably require that the licensed algorithm's source not be distributed.
Even worse: this license covers encryption. Most data is copyrighted by default, even if you don't include the circle-"C", therefor: it falls under the DMCA's prohibition on unlicensed decryption; it will be a criminal offense to even discuss compatible software.
What if Microsoft were to preemptively change the html or ftp protocols likewise? Microsoft customers are forced to follow; the standard rules of competition won't apply. Microsoft can end up owning every port and protocol on the net as their proprietary IP.
Microsoft will use our maligned IP laws to kill any Open Source project that attempts to be compatible. The Antitrust laws have failed; Microsoftis on a shooting spree. Samba will be first. Wine and compatible word processors and office suites will be next. But, they won't stop until every client and server is a Windows machine running MS applications.
If you don't think Ballmer is that ruthlessly competitive, then you haven't been watching him.
When I die, please cast my ashes upon Bill Gates -- for once, make him clean up after me!
IBM vigorously pursued any clone maker that didn't very carefully reverse-engineer their BIOS using clean room techniques. At least until very late in the game, you couldn't buy a BIOS license from IBM; if they called you on the carpet, the only course open was to revise your BIOS until IBM was happy with it.
In addition to threats of litigation against US clone makers, they also enlisted the US Customs service to impound shipments of PCs entering the US that IBM claimed had infringing BIOSes. In fact, IBM gave software to Customs to allow them to test for infringement -- if the IBM software said the BIOS was "too similar", the PCs were assumed guilty until the importer could prove their innocence. Several offshore clone companies died this way.
Phoenix and other BIOS companies that developed clean room BIOSes were the direct result of this, as was a very profitable business by National Software Testing Labs and others to do BIOS infringement evaluations. Lots and lots of money was spent to come up BIOS code that was both compatible and non-infringing. Writing assembly language interrupt routines that acted precisely like IBM's, but looked sufficiently different to avoid infringement claims was about the least pleasant programming task imaginable.
The Purple Book may have made clones possible, but IBM's copyright enforcement added lots of unnecessary cost to the clone market, without even adding much to IBM's bottom line (the money went to the reverse engineering industry, not to IBM).
Actually as I remember it Rod Canion (Spelling) tried to get his employer at the time (Texas Instruments) to sign off on his grand scheme to steal some of the thunder away from IBM by reverse engineering the BIOS an creating this thing called a PC "Clone". TI summarily dissed the idea since they knew there was no $$ in a consumer PC (Just look how poorly their TI 99-4A did) so Rod set about to do it himself (With $$ borrowed from the TI CU) in his garage. He created this company with a really funny name (Compaq) and then found someone to reverse engineer an IBM PC machines BIOS. Once the person was finished creating a Spec for the BIOS he was paid and sent on his way. Compaq then just used the Spec they had to create a BIOS that was compatible with IBM's. And that, junior, is how babies are born!
"A person is smart. People are dumb, panicky dangerous animals and you know it." - K
Actually, MCA wasn't all that horrific as far as the concept went. Don't forget that the PC was IBM's first "open" system, and that was only for expediency to get it to market in a year. IBM was still very much a big iron company, and thought that way. You could always be sure that IBM components would play well together, and MCA continued this concept into the PC world.
However, the fact that patenting MCA didn't improve IBM's share of the PC market should be a lesson in the advantages of truly open technologies over proprietary ones.
(And yes, my first Linux install was Slackware on a MCA PS/2)