Linux and Stone Soup
copyright 1997, Rex Ballard
Metaphors of the Linux and Microsoft Paradigms
article <firstname.lastname@example.org>, "Mike Palmer"
wrote: > mlw wrote in message > news:3832AB77.459057E0@mohawksoft.com...
> Interesting thread you've started here, Mark... >I'll state up front that I disagree with you, > but you're always interesting, and the debate is >almost always interesting...
I'm not sure I agree with all of Marks posting, but it does bring up some issues.
> > Linux and GNU is like the old days in a way.
Actually, Linux is copyrighted and published under the same General Public License used by GNU. In fact, it was first posted to an FSF site. All updates and enhancements since that first publication are protected by the GPL.
Linux is like "Stone Soup". The old lady begs door to door, and no one can give her anything. Finally she goes to a public area and puts a rock in a pot of boiling water and begins stirring it. Eventually, the curious come by, and they begin adding igrediants - carrots, peas, potatos, beef, cheese, and so on. Eventually the entire town can feast on this pot of soup made from scraps that would have spoiled. The same scraps that they couldn't give the old lady earlier.
The fable gives an insight into human nature. When the lady begged door to door, she offered nothing of value in return for what was being asked. The people being approached were threatened by this intruder. They had everything to lose, and nothing to gain.
When the lady began cooking the rock, it piqued people's curiosity. They wanted to share in this unique experience and thus were willing to contribute.
Linux was the Stone in Stone Soup
What Linux Torvalds did was start "cooking a rock". The original 10,000 line kernel was horrible. It made MS-DOS look sophisticated. If you didn't want to reformat your hard drive (which at that time cost nearly $1000), you had to swap your way through 3 floppies to get bash, sed, grep, awk, vi, gcc, and a few other simple utilities.
But people wanted to participate in this unique experience. They were learning as they developed kernel features, drivers, and libraries. The contributors, and even the promoters knew that they were getting value and were willing to contribute value. The result is literally a "soup" we now know as "the linux distribution".
Even today, much of the success of the Linux distribution is attributable to the fact that users who take the time to really experience Linux (as opposed to a poorly planned minimal effort) feel that they are getting much more than they paid for and are willing to contribute back in some way.
Result: $30,000 Workstation for $530
Linux is the $30 operating system that turns your $500 PC into a $30,000 workstation. To be fair, you can pay much more than $30 (up to $1500/user/year for top quality commercial support), and you can spend much more on a PC or Workstation (an Alpha or Power PC workstation can cost as much as $3000), and you can buy workstations from UNIX vendors for as little as $5000, but since 1993, Linux has delivered the power and performance of workstations that sold for several thousand dollars only a year or two earlier.
> > There is a software community, writing and > > developing software for the use of the community.And nobody's doing it for nothing. Some come in, eat a big bowl, and then start doing the dishes. Other's are busy getting the next bag of potatos (new software).
> > Hell we all like to make money, but at what cost?
Software Cornicopia - the Programmers Fantasy
The important thing is that Linux shatters the myth of the "software cornicopia". For years, I've watched reasonable men bet fortunes on the fantasy that they would create a "killer program like MS-DOS" and "become rich and famous like Bill Gates".
Bill Gates didn't just call up MITS, say "Here's a BASIC Interpreter, give me $100 billion", and collect his prize. Bill Gates III owed much of his success to Bill Gates II, a very successful lawyer who helped "Trey" craft the clever contracts that were crucial to hog-tying customers and punishing the disloyal.
Gates - Master of negotiating Aiki-Jitsu
Aiki-Jitsu is the combat form of Aikido used against soldiers. The aim is to redirect the force of the opponent, distract him, and then use his own force to kill him or critically injure him. For example, as the soldier would attack, the master could make it appear that he was just pushing the hand away, and at the last minute, would pull the wrist backward, while removing the eyes, and allowing the soldier to roll as he fell, but not letting go of his arm (thus dislocating the wrist, shoulder, and elbo with little more than a twist of the wrist. In a matter of about 3 seconds, an sword bearing soldier became a blind cripple.
The key to this strategy is appearing to be as helpless and defenseless as possible - encouraging the attacker to charge with as much force as possible, appearing to be amused are frightened (to prevent premature tensing) and to complete the disabling maneuver before the attacker actually realizes that he is about to kill or cripple himself.
Gates was very good at letting the customer think that a "bear trap" clause was "just a formality, standard practice", and usually in a whining voice that made you think it wasn't worth the trouble to go back for another round of drafts for. Gates often pretended to be weak in order to get the terms he wanted. In fact, he was a master negotiator, and extremely effective at making killer clauses that could be leveraged for massive gain look like "little nits".
Two of the most brilliant examples were when Gates asked for the right to set price tiers and the right to continue to enhance his core product. Keep in mind that at the time, the Judge had refused to accept a previous, more lenient agreement, was prepared to "throw the book at him", and had pretty much left Bill Gates hanging out to dry. At the time, it looked like Microsoft would be going out of the applications business.
By making the price tiers seem like a trivial matter, the DOJ didn't bother to work out the specific details of those price tiers. As a result, Microsoft set the prices in such a way that it was cheaper to buy more than you needed than it was to buy less than you needed.
By making "enhancements" seem like a minor point, the DOJ didn't bother to specify which enhancements. As a result, Microsoft was able to include all of the DLLs and OLE controls of Microsoft Office into Windows, and provide "Read-Only" access to Microsoft's proprietary protocols, along with "bare bones" implementations that assured that Microsoft's proprietary protocols, formats, and applications would be indespensible.
Another perfect illustration was when he minimized the desire to sell the software to other makers. He let IBM assume that he meant two-bit players who might sell 1-2% of the market for which Microsoft might get an extra few grand. Had IBM realized that Gates planned to sell "IBM's Operating System" to several key players in exchange for what amounted to 20% of the retail price on terms that allowed them to take 80% of the market from IBM, I don't think they would have agreed.
Planning into a distant future
Microsoft still depends on the fact that most people can't read Bill Gate's mind, and those who do have a clue are investing wisely and keeping quiet.
The average CEO plans no more than 5 years ahead, with most of his attention on the current quarter.
The average politician plans about two years (to the next national or state election) with most of his attention on the current poll.
Bill Gates plans 10-20 years ahead, with most of his attention on the collapse of his most threatening competitor.
UNIX - in the making since 1897 to be completed in 2097
There's only one problem. The people involved in the Linux/UNIX project planned as much as 50 years ahead. UNIX systems require a minor modification prior to the year 2038, to correct dates that are accurate to the year 0001 (including the transition from Cesearean to Gregorian Calendar). Alexander Graham Bell dreamed of a time when the deaf and the disabled could communicate as well as the hearing enabled. He saw the vision of the Internet in 1900, and that vision is being fulfilled 100 years later.
The "Desciples of Bell" managed and handled UNIX, creating the infrastructure, including a carefully managed relationship with the academic, government, and research communities. Eventually, many of them left AT&T to take their places at companies like MCI, Sprint, GTE, and all of those little ISPs and IAPs.
When Bill Gates was announcing in an interview with a major computer magazine how he would go from leader of the largest software company (at $100 million/year) to "World Domination" within 10 years, there were members of the UNIX community setting the stage to topple him in 20.
Microsoft in a Competitive Market
As strange as it may seem, there is little desire to destroy Microsoft. The information and communication industry (and you thought it was about computers) is huge, and there's plenty of room for everyone, including Microsoft. Unfortunately, the same problems that caused the AT&T monopoly and the IBM monopoly to hamper the growth of the industry are happening at Microsoft. AT&T resisted the PC, resisted modems, and struggled to retain total control of "the wire" (fibre) long after it was possible to share it. IBM resisted open standards, published formats, and private control of information long after it was possible to provide it.
The 21st Century Vision
The vision is one where everyone has access to information and communication, in a form that is easy to use and fully functional. When people can communicate, they can resolve differences without resorting to violence. When they can communicate, they can create opportunity where there was once poverty in the presence of a population willing to contribute generously.
Microsoft made a huge contribution toward this dream. They created computers that were simple to use, relatively inexpensive, and fully functional (if not terribly reliable). No one has the right to drive them "out of the market".
Could the Dream become a Nightmere?
Today, Microsoft has become the obsticle. Gates has become so obsessed with protecting his "99%" that he has attempted to stifle the innovation of others. It's likely that many of those who now hold top positions at Microsoft were drawn by the possibility of weilding the same type of power as the henchmen of Capone, Hitler, Napolean, Gengis Kahn, or Julius Ceasar. And when you consider the accomplishements of Bill Gates, you must acknowledge that he has accomplished no less than any of these, without the bloodshed, the violence, and the brutality of his predecessors.
But each great conquerer wields his own brand of terror. Ceasar had the Roman Legion and Crucifixion. Gengis Kahn had the Mongol "hordes" and the massacre of entire cities. Napolean had the great armies and the Guillotine. Hitler had the SA, the SS, and the Death Camps.
The Ultimate Terrorist
Microsoft wields a different, more subtle type of terror. Microsoft's leaders understand that in the American Culture, the American will fight tyranny violently, but they will protect security and economic stability covertly. To the leaders of American Business, their very identity and existance is based on the success or failure of their companies, often on a quarter by quarter basis. The reward for success is wealth, power, international recognition, and sanctification. A successful CEO or Manager is treated like a prince, a lord, or a duke.
But the penalty for failure is worse than death to the American Manager. He can be publicly humiliated, asked for his resignation, deprived of luxuries and benefits to which he has become accustomed, and eventually "given the opportunity to offer his resignation". This is only the beginning. The loss of prestige, power and income can result in loss of opportunity, loss of income, and demand for immediate repayment of debts. The once great prince becomes a pauper.
Typically, the greatest fear of the CEO is "Skid Row", because it is within the realm of possibility. The technician can ply his trade anywhere, he needs only his tools and his knowledge. But the manager depends on the respect and obedience of those who believe that by making him successful, they will become successful. When a leader fails, even the best efforts of his subordinates cannot prevent the loss of jobs, of economic hardship. In American cultures, leaders are directly responsible to those they serve, they lead by consent of the governed. When that consent is withdrawn, they are no longer leaders.
In Europe, Asia, or Africa, a leader's career is ended in a relatively quick, relatively merciful, relatively dignified death, either suicide or execution. The body may be abused after the death, but the dictator no longer cares.
The 20 year torture
In the United States and Canada, a leader's career is ended in a relatively unpleasant way. After being stripped of all rank and power, he is publicly villified, humiliated, and dispised. The only people who will even associate with him are drug addicts, alcoholics, and $20 hookers. Often, his wife and children will leave him, taking everything they can get, and then demanding monthly payments amounting to half the income he used to make. When it's all gone, he isn't even worth the effort to put him in jail. Eventually, they may die of alcoholism, drugs, or AIDS, others just end up bitter old men, pan-handling in the streets of cities like New York, trying to make just enough to buy some coffee and a bottle of cheap wine so they can go back to sleep in the subway tunnels.
The New Leadership of "Trusted Servants"
The only leaders who are immune to this type of threat are those who have already "been there, done that", and "came back from the dead". To them, there isn't much fear of failure. They succeed because they understand that the only thing that matters is that OTHERS benifit from their leadership. Many will even avoid the trappings of power. Richard Stallman controls a software base nearly twice the size of Microsoft's entire code base, and yet he lives a rather bohemian lifestyle. Larry Wall and some of the other Open Source leaders control the software base used in the Internet - which has generated nearly $1 trillion for it's users, and yet they choose to support Open Source as modest players on a huge team.
Even Bob Young is incredibly modest and moderate. He remembers the days when he went to New York and couldn't pay for dinner until he sold 50 Red Hat CD-ROMS at Javitz Center. Those are the stories he like to tell.
In fact, most of the leaders of the Linux Movement are merely stirring a big pot of Stone Soup, offering contributors and customers alike the opporunity to enjoy a huge feast they created themselves.
Patents, Trademarks, and Copyrights
> > This whole software > > patent nightmare is making development of new software almost > > impossible.Software patents have always been problematic. First, it is very easy to intuitively derive many systems which have been patented. It's like trying to patent "Einstein's watch" by describing the movement of the hands. The problem is that unless the mechanism is exposed, you can't tell whether your watch is the same as your competitor's. One may be a Rolex Mechanical watch with springs and gears, and flywheels, while the other is a Casio with a simple stepper motor under control of a quartz crystal.
More often, software patents are nothing more than protection of a "trademark". For example, someone develops a particular format for storing or sending information. They can't copyright it because the copyright only protects their expression of the format. A different expression (for example a different programming language, notation, or algorythm) produces the same format, but it is a unique expression.
The Open Source community was founded on the premise that only the creator of the expression owns his creative efforts, that the format used by the word processor does not entitle the author of the word processor to control distribution of the original work.
Microsoft has exactly the opposite view. In their license agreements, if you publish a document in Microsoft Word format, for example, you agree not to reverse-engineer the document to expose the format. Those who do have access to the format are required to sign nondisclosures as well, which often supercede previous knowledge. For example, if you knew HTML before you accepted the terms of the Microsoft Developer's Network Licence Agreement, any future disclosures of HTML format require approval from Microsoft.
Any copyright, patent, or license is only as valuable as it's enforcement. Microsoft makes a point of routinely exposing and enforcing it's copyrights, licenses and patents, but usually either very covertly (subject to even more nondisclosure agreements), or overtly against defenseless defendents such as PC stores that are teetering on the verge of bankruptcy anyway. In most cases, the issue is settled out of court, usually for both financial and non-monetary consessions, for example 20% of the company, the exclusive right to all software developed by employees (which is why so many companies now use contractors or consultants instead of their own employees to develop software), and control of all future press releases related to Microsoft. Often, Microsoft chooses it's targets with these aims in mind.
Eventually, software intellectual rights become a liability. During the 20 year life of a patent, it's very easy for competitors to define an equally useful standard that does not infringe on the original patent. Unfortunately, the patent holders will often overextend themselves trying to protect and perpetuate the patented or copyrighted product.
IBM promoted System 370 over UNIX for years, and it eventually became a crisis in 1992. By the time they began implementing the UNIX paradigms in OS/390 the stock was tumbling horribly. New leadership gave up the attachment to "Developed In House" brands and started focusing on what the customer wanted and needed. Now, IBM is much larger than it was before.
DEC promoted VMS long after UNIX had taken control of the market. Eventually, under the leadership of Compaq executives, priorities were shifted to the customer's goals and needs. If Compaq would give up it's attachment to NT... magic could happen.
Why is Microsoft so Agressive (a silly theory)
> > Nothing I have seen patented lately has been anything but > > common sense. Does Microsoft have to crush every last > > bit of competition it can?
Bill is perceived as a "Nerd". No matter how many billions he makes, he is treated like a Nerd by most of the business circles he deals with. To his face, they will be polite and even respectful, but behind his back, corporate leaders are often dominated by the "jocks" who seem to dislike Nerds. Even when Bill could "squish 'em like a grape", bankers, lawyers, brokers, and politicians treat him like the kid with the horn rimmed glasses held together with adhesive tape and the big slide rule or calculator in his pocket - to them, it's a joke.
Ballmer still makes jokes about it. But Bill isn't really laughing. Bill uses that Nerd image to win the support of IT experts, who are often a bit "nerdy" themselves. Many "nerds and geeks", or those who identify with that term see Bill as their big hero. He's the nerd who became the richest man on earth.
But the "new nerds" notice that Bill isn't helping THEM much. They work 60 hour weeks and then go home to program for Microsoft Windows thinking that they are going to "get rich quick", only to realize, often after years of contributing software to Microsoft and Microsoft supported causes, that there's no cheese down that tunnel.
Linux - few promises, just results
Linux starts with a different premise. They start with "hey, we'll start out by giving you as much code as you want, and when you contribute, your name goes on the credit lists. We'll start out by giving you more than Microsoft ever gave you, and give you the opportunity to profit by selling support.
The model works pretty well. Red Hat makes pretty good money putting together a box for $20-$100 bucks that usually requires only a phone call or two to sort out. LinuxCare makes quite a bit more selling 24/7 support to ISPs and other "hard core" users. IBM even gets into the game, providing consulting services, system integration, and user support for Linux using customers. Actually, there are several outstanding Linux support providers.
Winner Take All, and other Myths
> The nature of business is to win.
At what cost? Every game has rules. Even war has rules.
When you play football, your aren't allowed to pull your 9mm and shoot the quarteback in the head.
When you play war, you have a gun, the other guy has a gun, but if one of you starts using nerve gas or anthrax spores, the other might start using neuclear warheads. Since the turn of this century it has been within our power to wipe out the entire population of the planet, the ultimate outcome of abolishing all of the rules.
There was once a time, in the 1930s, when business leaders dealt with competitors using a Thompson submachine gun. In Chicago, Al Capone gained a monopoly in several business lines which he defended "simply providing pleasure that customers want". The government decided that they wouldn't make the rules for selling alchohol (by making it illegal). As a result, there were no rules and the winner was the one who could get the guys who had machine guns past the guys who had machine guns to turn the leaders (who didn't have machine guns) into pegboards (lots of little holes)...
There was a time when monopolistic railroads and steel mills dealt with union leaders by burning crosses on their doorsteps, hanging them in the middle of the night, or just shooting them with a machine gun or a sawed off shotgun. Eventually the government had to step in and make laws for both the Unions AND the Businesses.
The 21st Century Tactics and Strategies
In "cyber wars" the weapons are information, communication, and control of both. In the absense of any regulation Microsoft could eventually demand the "right" to add "extensions" that can tap your phone (interphone), read your mail (e-mail), audit your books (MS-Money), expose your sexual preferences and practices (following your browser history), or even send mail on your behalf and help you clean up you hard drive by copying the contents of your hard drive to Microsoft's on-line data archive (where intellectual property rights could be pillaged without leaving evidence).
When Hitler's SS, or Stalin's KGB or even Orwell's fictional "thought police" attempted to invade privacy, nationalize private information, and blackmail business and political leaders, the world objected.
When Microsoft offers to provide the same "benefits" for the exclusive gain of Bill Gates, they pay a premium price for the priviledge of letting him have it.
> I seem to remember a few years ago that > Larry Ellison and Scott McNeally were > touting the network computer, or > information appliance, or some such,Which was actually Linux in disguise. It's interesting that Linux was one of the first widely used web servers, pop servers, mail servers, and news servers, both inside and outside the firewall.
Microsoft was especially fond of "taking posession". The earliest "thin clients" were PCs that couldn't be upgraded to Windows 95 but could be converted to Linux. Linux didn't have many applications of it's own, but it could run a web browser and a decent implementation of Java. Eventually Microsoft killed the NC by porting IE 4 to Windows 3.1. Eventually, the prices of PCs dropped to the prices originally planned for the NCs. The incremental cost of the hard drive coupled with poor network connectivity and management pretty much killed demand.
> and predicting that it would kill Microsoft.
Essentially, the Linux computer is not much more than a Network Computer with it's own built-in server. Even the GNOME and KDE use components in which simple GUI clients are connected to powerful servers using CORBA.
UNIX has always succeeded by attacking the underbelly. The entry-level price points were lower than their competitors. To employees, knowing that budget cuts were coming, the choice of a low-cost UNIX system or a high-cost proprietary system, followed by the higher priced support, UNIX was a nice way to save a few jobs, including their own.
Until Windows 95, Microsoft had such a low underbelly, that it was nearly impossible to get under it. But with NT, Microsoft created the opening for the underbelly. Microsoft was selling $1000 NT server licenses, and $400 workstation licenses, along with deliberately confusing $30 client access licenses. Many managers actually think that if they buy NT Server, that they have permission to install NT workstation on 10 machines - in reality, they still have to pay $400/machine AND $36 for each machine that ACCESSES that machine (whether Windows 3.1, Windows 95, Windows 98, or Windows NT). I've known several managers at Fortune 500 companies that nearly committed suicide because they discovered after installing thousands of workstations, that they owed Microsoft millions of dollars.
The confront of this reality has been a wake-up call for many IT managers. After getting hit for $ millions to be upgraded to NT 3.5x, then upgrade to 95, then upgraded to Office 95 then upgraded to NT 4.0 then upgraded to Office 97, then upgraded to Windows 98 then upgraded to Office 2000 then upgraded to Windows 2000 (whenever it comes out) then upgraded to Millenium (2-3 years from now?)... they are saying "enough is enough, tell us about this Linux thing".
Microsoft is promising that Windows 2000 will offer the "feature" of "automatically upgrading" your system. The folks who have had so much fun with NT Service Pack 2 on Cyrix systems, or Service Pack 4 with Notes 4.0 and Netscape 4.01 can't wait. Just think, Microsoft will "auto-upgrade" all of their PCs, at 3:00 A.M. local time, and everyone will come into work only to discover that the CRM, Contact Management System, and ERP system have just been blown out of the water - and the market opens in 5 minutes. They ported to Linux - so they're dead.
The first time this happens of course, Microsoft will watch millions of customers pull the plug on Windows 2000 the same way they pulled the plug on NT 3.5. And this time, instead of going back to the predecessor Microsoft product, they will be looking at Linux. I'm not looking forward to the first Auto-download of SP2 for Windows 2000 - when Microsoft advocates think it's safe to go in the water, and Microsoft blows them out of the water... messy--very messy.
Never Lose Sight of the Customers
> Then I remember how fast DEC went > from having a virtual lock on > the minicomputer market to having > virtually nothing, and I can see why > Microsoft feels threatened.Part of the problem was that DEC was selling lots of Vaxen with preinstalled VMS which were being converted to BSD UNIX systems. DEC kept claiming that they had this huge VMS base, while the reality was that only a small percentage of VMS systems were still installed.
Eventually, DEC began sending it's service REPs to install "Patches", which included removing any non-VMS operating system (UNIX) and reinstalling a brand new copy of VMS. Fortunately, the Hard Drives came with removable platters and when the REP would schedule an upgrade the Admins would install the VMS pack. The REPs didn't even notice that there was 2 gig of storage (6 washing machine size 300 meg hard drives) that didn't show up in the VMS allocation table, but what mattered was that they could legally say that the system was runninng VMS when they got there.
Unfortunately, DEC eventually started believing it's own press releases. Within DEC, a small group of Berkely Graduates offered to make a DEC version of BSD (Ultrix) but the corporate politics was so intense that DEC ignored UNIX and pushed VMS. That was 1987.
When Sun and HP started really supporting the server market, and created their own cluster technologies, DEC couldn't stop the bleeding fast enough. The PC division kept it afloat, and the Alpha kept investors interested, but DEC eventually began cutting almost the entire VMS division (who migrated to Microsoft and NT development). Microsoft eventually paid DEC and officially licensed NT, and agreed to port to the Alpha, but it was too late. It was actually even a bit messy for Compaq.
> If you're going to win, especially in a > marketplace where technology changes can > kill your multi-billion dollar > business in a couple years,UNIX almost killed IBM. IBM had fought UNIX for years, largely because they feared that AT&T could become a rival in the computer industry. In 1991, the "bet the farm on MVS 4.0" and lost the bet. John Akers had played to politics of the insider organization and ignored UNIX in a big way. They poured more than they should have into MVS, OS/400 and OS/2 development and practically starved the Austin RS/6000 departement.
The problem was pretty simple, MVS 4.0 required a whole series of upgrades, all at the same time. The price tag to upgrade a $6 million 3090/ES system into an MVS 4.0 was almost $6 million. I remember suggesting an 8 processor Pyramid that cost $250k and being concerned that my employer couldn't afford it. What I didn't know was that unless we moved to the Pyramid, we were going to have to pay $30 million in upgrade costs. We got the Pyramid, a Sun/6, AND a Sequent and orders to port everything we could. We weren't the only ones. IBM's stock dropped 50% in two months. That was 1991.
When Lou Gerstner took over, he shifted the whole focus to the customer. It took a few years, but eventually, the IBM line and staff realized that there were a whole bunch of customers and potential customers who were using UNIX.
Today, UNIX, and the UNIX paradigm is used across the entire IBM product line from OS/2 to AIX to OS/390. These platforms even support compilation of Open Source source code.
When NT Workstation sales fell very short of expectations, Microsoft
decided to make NT into a server. They tuned the system a bit
differently and said "you can use NT just like UNIX".
The problem with a server is that you don't want the success of your
operation depending on whether your 3rd shift operator wakes up
from his 3:00 A.M. nap to push an "idiot button" that will start
reorgs or backups or services. If he sleeps a little too long,
you have a production system, with a reorg 85% complete 15 minutes
before market open, and you have to pull the plug to the Internet
to keep your database from blowing itself to kingdom come.
ADMINS started using PERL because the Borland Compiler could
compile ANSII C well enough to port GCC to NT. The GCC could be
used to compile PERL, and the administrators had a language that
had SOME of the scripting capabilities of UNIX. They had to
disable a whole bunch of functions because NT didn't support them,
but at least you could parse some text, start some jobs, and
run a scheduler so that you could get a good night's sleep.
The big probem for Microsoft was that PERL also supported a number
of portable APIs that made it possible to create desktop applications
for both NT and UNIX (and Linux). You could use TK, GTK, or QT
or just use PERL and NCSA (later Apache). Microsoft could lose
all NT Desktop users to UNIX or Linux.
Microsoft funded development of PERL extensions that prevented
or eliminated the need for portable APIs. They did the same
thing with SoftwareAG and DCOM. Microsoft wanted to use DCOM
to keep CORBA off of Windows, so they funded an effort to
port a subset of DCOM to UNIX. When SoftwareAG started
weaving DCOM and CORBA together so that they were orthagonal
(CORBA objects could call NT DCOM services and NT DCOM clients
could call CORBA services), Microsoft "pulled the plug" and
took control of DCOM for UNIX development.
The biggest threat to Microsoft is the market-wide acceptance of
any API and/or toolset that would make it possible to create
identical applications for both UNIX/Linux and Microsoft Windows.
Microsoft is particularly concerned about Linux because widespread
acceptance of Linux would eliminate Microsoft's bargaining leverage.
While most folks in the industry just try to guess at the approximate
size of the Linux Market, Microsoft counts very *nix box that touches
MSN, microsoft.com, msnbc.com, or any of Microsoft's other sites. And
they ussue cookies from the first page visited, just in case you
don't want to fill out one of their nice forms. Microsoft knows
EXACTLY how many machines have had Linux loaded onto them at least
enough to be used as a web browser (very bad news for Microsoft).
Of course, Netscape and AOL do the same thing with their sites.
Microsoft doesn't want anyone to know how many Linux users there
really are (actually ever have been), and Netscape doesn't want
anybody to know either. Netscape wants a favorable judgement
that opens the market to competitors such as Linux and prevents
Microsoft from closing the holes. Microsoft wants the OEMs to
think that Linux isn't a big enough market to risk forfeiting
even secondary markets like Windows 2000.
If Windows 2000 were the successor to Windows 98 ($100 retail
price tag, installed on every OEM machine...) the OEMs might
be less bold about adopting Linux. Microsoft "bet the farm"
on Windows 2000 and it looks like the hardware requirements
are such that it won't be able to penetrate the markets that
are now nearly saturated (Business PCs, high end home PCs...).
When Microsoft announced "Millenium", it admitted it lost the
bet. Windows 98 isn't a premium product and didn't drive
the market in any big way (nobody's junking Windows 95 machines
to get Windows 98 machines and most Windows 95 machines don't
survive Windows 98 "upgrades" very well.
Microsoft's "trump card" was the "all or nothing deal", which is
what the DOJ is trying to prevent. The strategy here is that
unless OEMs abandon the Linux market, Microsoft will revoke
their Windows licenses (or refuse to renew). They must play
this card before the Linux market becomes measurably large
enough to make the threat moot. Desparate people do desparate
> -- Mike --
Who put the Overalls in Mrs. Willie's Chowder!
> It all depends. Microsoft apparently does see a benefit to some Gnu
> software - they fund ActiveState's development of Perl for NT, for example.
> For Microsoft, it enhances NT, even though it helps Gnu at the same time.
This may seem very generous, but it has alterior motives. When
NT 3.x first came out, there wasn't going to be a scripting
language. The command.com would still be there, but everybody
would do everything with Visual C++ or using Microsoft's
commercial applications. NT was originally intended as
a single user workstation not as a server.
"SOUP"?, I thought you meant "SOAP"!
> In other areas, like compilers, it would
> be more difficult to imagine Microsoft
> funding development - they are a competitor.
When NT Workstation sales fell very short of expectations, Microsoft decided to make NT into a server. They tuned the system a bit differently and said "you can use NT just like UNIX".
The problem with a server is that you don't want the success of your operation depending on whether your 3rd shift operator wakes up from his 3:00 A.M. nap to push an "idiot button" that will start reorgs or backups or services. If he sleeps a little too long, you have a production system, with a reorg 85% complete 15 minutes before market open, and you have to pull the plug to the Internet to keep your database from blowing itself to kingdom come.
ADMINS started using PERL because the Borland Compiler could compile ANSII C well enough to port GCC to NT. The GCC could be used to compile PERL, and the administrators had a language that had SOME of the scripting capabilities of UNIX. They had to disable a whole bunch of functions because NT didn't support them, but at least you could parse some text, start some jobs, and run a scheduler so that you could get a good night's sleep.
The big probem for Microsoft was that PERL also supported a number of portable APIs that made it possible to create desktop applications for both NT and UNIX (and Linux). You could use TK, GTK, or QT or just use PERL and NCSA (later Apache). Microsoft could lose all NT Desktop users to UNIX or Linux.
Microsoft funded development of PERL extensions that prevented or eliminated the need for portable APIs. They did the same thing with SoftwareAG and DCOM. Microsoft wanted to use DCOM to keep CORBA off of Windows, so they funded an effort to port a subset of DCOM to UNIX. When SoftwareAG started weaving DCOM and CORBA together so that they were orthagonal (CORBA objects could call NT DCOM services and NT DCOM clients could call CORBA services), Microsoft "pulled the plug" and took control of DCOM for UNIX development.
The biggest threat to Microsoft is the market-wide acceptance of any API and/or toolset that would make it possible to create identical applications for both UNIX/Linux and Microsoft Windows.
Microsoft is particularly concerned about Linux because widespread acceptance of Linux would eliminate Microsoft's bargaining leverage. While most folks in the industry just try to guess at the approximate size of the Linux Market, Microsoft counts very *nix box that touches MSN, microsoft.com, msnbc.com, or any of Microsoft's other sites. And they ussue cookies from the first page visited, just in case you don't want to fill out one of their nice forms. Microsoft knows EXACTLY how many machines have had Linux loaded onto them at least enough to be used as a web browser (very bad news for Microsoft).
Of course, Netscape and AOL do the same thing with their sites.
Microsoft doesn't want anyone to know how many Linux users there really are (actually ever have been), and Netscape doesn't want anybody to know either. Netscape wants a favorable judgement that opens the market to competitors such as Linux and prevents Microsoft from closing the holes. Microsoft wants the OEMs to think that Linux isn't a big enough market to risk forfeiting even secondary markets like Windows 2000.
If Windows 2000 were the successor to Windows 98 ($100 retail price tag, installed on every OEM machine...) the OEMs might be less bold about adopting Linux. Microsoft "bet the farm" on Windows 2000 and it looks like the hardware requirements are such that it won't be able to penetrate the markets that are now nearly saturated (Business PCs, high end home PCs...).
When Microsoft announced "Millenium", it admitted it lost the bet. Windows 98 isn't a premium product and didn't drive the market in any big way (nobody's junking Windows 95 machines to get Windows 98 machines and most Windows 95 machines don't survive Windows 98 "upgrades" very well.
Microsoft's "trump card" was the "all or nothing deal", which is what the DOJ is trying to prevent. The strategy here is that unless OEMs abandon the Linux market, Microsoft will revoke their Windows licenses (or refuse to renew). They must play this card before the Linux market becomes measurably large enough to make the threat moot. Desparate people do desparate things.
> -- Mike --