 Michael and Rebecca Wong |
 |
Michael and Rebecca Wong |
|
The history of Microsoft and IBM is long and tortured. Interestingly enough, there are some parallels in the historical behaviour of these two giants.
Does anyone remember the good old days of IBM's hegemony over the computer industry? They used a combination of proprietary hardware and software to control their mainframe and minicomputer customers. Price/performance ratios were irrelevant because they used the spectre of incompatibility and impossible migrations to terrify their customers into staying put. They routinely introduced new products and then forced their customers to upgrade, because old software would no longer be supported and new software would only run on new hardware. This situation is endemic to the computer industry as a whole, but the crucial difference between the general computer upgrade cycle and IBM's monopoly was that IBM held the reins on both sides of the equation: hardware and software. They could use software as leverage to force people to buy new hardware, and vice versa. This anti-competitive practice was extremely harmful to consumers and led to an antitrust suit, which IBM's lawyers dragged out for so long that IBM lost its throne to Microsoft before it could be resolved, thus making the entire trial moot.
Of course, that leads to the sordid tale of how IBM gave away its market dominance. The PC heralded the beginning of the end, simply because it didn't fit into IBM's business model. They offered it as a triviality, farming out the operating system to a little-known company called Microsoft. This was a strange business decision for a company that had built its success on control, but they honestly didn't think the PC would ever amount to anything. But IBM didn't truly dig its own grave in the late 1980's, when it decided (too late) that the PC market had grown to the point that it should be taken seriously. By this time, competing PC manufacturers such as Compaq were selling machines that outraced IBM's machines, and for less money. Everybody seemed to be getting in on the act, and IBM management decided that it was time to rein in this market before it got away from them. Their solution came as a two-pronged approach: the OS/2 operating system and the PS/2 computer with Micro Channel. Both had their strengths and weaknesses, and both failed.
The failure of OS/2
The first half of IBM's plan for world domination was a stable, powerful, and proprietary operating system: OS/2. Technologically, it was a superior product; in fact, ten years of Windows development still failed to produce an operating system that was as stable as OS/2. But IBM designed it exclusively as a workplace OS, so it wasn't 100% DOS compatible and it wasn't very good for games. This was a mistake (and it was totally, tragically unnecessary; see the Postscript at the end of this page). Compare that to the way people warmed up to Windows 95, which had special features for gaming and which offered "MS-DOS mode" for 100% DOS compatibility. Skittish business users were reluctant to buy something that wasn't guaranteed to be 100% compatible, and home users ignored it completely, as its designers had ignored them. Worse yet, there were no "killer apps" to entice users to OS/2. And as the final nail in OS/2's coffin, Bill Gates abruptly stabbed IBM in the back by pulling out of OS/2 development and aggressively pushing Windows instead.
The combination of low cost, guaranteed 100% DOS compatibility and a growing base of apps was a magic one for Windows, which began to dominate the business desktop in spite of massive architectural flaws. OS/2 fell to the technologically inferior Windows glorified-shell because IBM forgot one of the critical rules of sales: ask the customer what he wants. OS/2 was a technologically superior operating system, but it solved problems that weren't yet on the average consumer's radar and it didn't address issues that the customer did worry about (such as guaranteed compatibility).
The failure of Micro Channel
The other half of IBM's plan for world domination was a proprietary peripheral interface for computers: Micro Channel. The Micro Channel Architecture, or MCA, was a replacement for the horrible ISA bus. It debuted in the PS/2, and it was designed to solve certain glaring technical problems in the PC platform while simultaneously eliminating competition in the hardware market. Like the OS/2 operating system, it was technologically advanced but it was doomed to failure.
The MCA failed because it was a truly closed standard. Not only was it proprietary and subject to licensing requirements, but it relied upon an unwieldy and expensive registration process. Anyone marketing an MCA card had to register it with IBM, because each device required a unique ID number and compatibility certification. Nobody could make Micro Channel machines without IBM's permission, and nobody could make Micro Channel cards without going through IBM for certification and registration. Welcome to the monopoly.
I was actually an employee of IBM in 1988, and I remember asking a marketing executive what their strategy was for dealing with "the Compaq threat," since Compaq was the undisputed leader of the so-called "IBM clones" at the time. Compaq and the other clone-makers were selling computers that outraced IBM's computers for much less money, and businesses were taking notice. Amazingly, he calmly replied that IBM didn't need a strategy! In his view, all clone-makers were doomed by the new proprietary MCA standard. He expected that all of the clones would either stay with ISA and fade away as the new standard took hold, or they would switch to MCA and pay royalties to IBM. Either way, IBM would win! Of course, history tells us otherwise. The PC-compatible industry responded to IBM's shot across the bow with a return salvo of its own: the EISA (Extended ISA) standard.
Nobody wanted to pay the outrageous prices that IBM was charging for MCA PS/2 machines, or the similarly outrageous prices being charged for MCA devices. When EISA appeared, the market stampeded toward the new open standard, and MCA was left to wither in the dust. The EISA bus was since supplanted by the VESA VL-bus and more recently, the AGP and PCI buses, all of which were open standards. There's a lesson here, if anyone's willing to listen. People do prefer open standards (note that "open standard" is not the same as "open source"). If given a chance on a level playing field, they'll take them over proprietary standards any day.
Microsoft: Destroyer of Standards
IBM attempted to take open standards and replace them with closed, proprietary standards. Microsoft doesn't emulate this pattern precisely, but as you might expect, Bill Gates has invented a sneakier, more devious way of accomplishing the same goal. Here is the Microsoft Modus Operandi:
If the market will bear it, replace open standards with proprietary Microsoft standards, as per IBM's old MO.
If consumers won't drop the open standard, then adopt it, but not before "enhancing" with proprietary Microsoft extensions. Flood the market with applications and tools that are dependent upon these extensions. Then, publicly declare that any tools based on the original open standard are inferior because they don't support the "enhancements". All of a sudden, an open standard has become corrupted into yet another piece of Microsoft property.
Every time Microsoft encounters a situation screaming for an open standard (or already serviced by an open standard), they either invent their own closed standard or they corrupt an existing one. They will only deviate from this course if the market forces them to do so. The history of their behaviour is interesting:
Instead of using the open POP mailbox standard, the first Windows mail implementations used the oddball Microsoft Mail standard. The market ignored Microsoft Mail so they capitulated and made Outlook Express, which can use POP or IMAP.
Instead of using the open SMTP mail transport standard, Microsoft invented the MAPI protocol for Exchange. They've aggressively pushed to have all Windows applications use MAPI to send mail instead of SMTP, even though the Internet still runs on SMTP. They even have the gall to call MAPI a "standard"! Once again, the market rejected their standard because SMTP servers worked fine.
Instead of using TCP/IP or even Novell's IPX, Windows tried to promote the ridiculous non-routable NetBEUI network protocol. Does anyone even remember this useless protocol any more? For many years, Microsoft tried to force users to adopt NetBEUI by making it the default protocol and causing Windows networks to be unreliable without it, but users wanted the Internet. As with POP/IMAP and MAPI, Microsoft eventually capitulated to TCP/IP.
Instead of using Java as-is, they tried to add Microsoft "extensions" to it, which would only work on Windows. This represented a wholesale corruption of the entire platform independent concept of Java, and it was obviously a deliberate attempt to corrupt Java into a Windows development language. Sun sued them to stop this behaviour and won in court, so Java works the same on all platforms.
Instead of using the open standard X.500 or helping to develop the open standard LDAP, Microsoft invented Active Directory for Win2000. Instead of using the open standard NIS for network information services, Microsoft invented the undocumented and proprietary NT domain protocol. This was successful, largely because it was not an attempt to monkey with the basic operation of the Internet. Active Directory is really just a Windows Domain function and is totally irrelevant to most daily computer operation, particularly for home users.
Instead of using LDAP for addressing, Microsoft uses Exchange. Outlook Express doesn't even have an option for LDAP. But the entire personal information management market took an unexpected turn when handheld devices essentially stole it away from PCs lock, stock, and barrel. Microsoft failed to control this market because buyers of handheld units simply didn't see any advantage to Microsoft products. Their standard practice of leveraging market share in one market to force it in another market was ineffective in this case.
Instead of using HTML as the open cross-platform document formatting standard that it was meant to be, Microsoft attempted to corrupt it into a Windows language by adding all sorts of Windows-specific "extensions" to it (those ActiveX and ASP extensions that the IE and IIS people are always bragging about). They did achieve limited success with this tactic, but the vast majority of websites on the Internet do not use ActiveX controls. The market looked at ActiveX and asked "what's in it for me?" Microsoft had no answer, and the initiative largely failed.
Instead of being content with standard E-mail content, they had to add the ability to execute Visual Basic scripts on Outlook Express: a "feature" that almost immediately became a gigantic security hole and which has since been heavily locked down. Once more, the Internet market rejected Microsoft's attempt to change the way things are done, and almost no one ever sends or receives E-mail with Visual Basic extensions any more.
Instead of using the widely accepted and robust OpenGL OpenGL standard, they created the proprietary Direct3D API. It took more than a half-dozen iterations of this API before it surpassed OpenGL's feature set. This was a success for Microsoft, and it was a major contributor to their dominance of the home computer gaming market.
As you can see, despite their reputation for unbridled success, they actually have a rather "hit and miss" record with their strategy of replacing open standards in favour of proprietary ones. And there is a pattern here: the Internet has consistently confounded their attempts at control. Their biggest success was Direct3D, which has nothing to do with the Internet.
Like IBM in the past, Microsoft has a tendency toward arrogance, and often assumes that it can simply dictate terms to the market. They often believe that they can create a new proprietary standard which offers no real benefit to anyone but themselves, and then get everyone to adopt it anyway, by leveraging their Windows market share. That's textbook anti-trust, but anti-trust law seems to be dead in America's pro-corporate culture right now.
I received an E-Mail on August 25, 2001 from Rich Heimlich (a name which may be familiar to longtime computer newsgroup surfers), relating some fascinating details of the OS/2 saga which I was not previously aware of, and which reveal some of the arrogance of Big Blue:
On the IBM OS/2 front and why it failed I can lend some real insight into that which supports your comments. I owned the oldest and most respected quality assurance company in the entertainment industry from 1983-1996. When Warp was in development IBM paid my company to test Warp against a suite of more than 400 games. I personally invested my time to test these products for them. When an overwhelming number of them failed, our company did some research and found out that a single driver was the main cause. It was VDMA.SYS or the Virtual DMA driver. The problem was it was keeping nearly every program from working with sound and either resulting in no sound or crashes due to the sound trying to work.
I took this to IBM and pointed it out to my main contact, a major exec there who took it right to the top brass and I mean to the top 3 people in the company short of the CEO. We had a phone conference where I suggested that they allow VDMA to be opened up so that games could function. Their response was that the games did function. I countered by saying, "Look, DOOM (the big game at the time) runs, but not without sound." They said, "So? It runs doesn't it?" I told them that DOOM without sound was like "Gone With the Wind" without sound. What's the point? I then asked if they preferred silent movies over "talkies" and of course they didn't get the correlation.
Finally, I pointed out that the sheer number of failing games was a problem and their true colors came out. My contact supported the view 100% but none of the brass did. They felt that games were not important and that allowing VDMA.SYS's security to compromised was too big a risk for an OS that was designed to be secure for business. I then told them Warp and OS/2, if it stayed that course, was done and they of course disagreed.
Then, within a matter of days, they fired my contact there (which was well covered by Will Zachmann of PC Magazine at the time) over the whole issue and the rest, as they say, is history.
Sad, isn't it?
Last changed: 2007/03/31
Continue to General Mailbag
Jump to:
This site was written using KWrite and OpenOffice, running on Mandriva Linux.