Phil Regnauld er franskmand og taler flydende dansk. Skriftligt dansk kniber det lidt med, så denne artikel er på engelsk. Phil er mest kendt for sine aktiviteter indenfor FreeeBSD i Danmark. (chlor)
Painting the corner
Feb 1999 (v.0.2)
1. IntroAs I look back on the events that have occured in the last 6 months in the software world, and try to imagine what the next half year will bring, I get the strong impression that we are on the turning point of many things.
To be schematic -- binary ? -- here are the issues:
Now, depending on your background as a computer professional -- or not -- each of these questions will mean more or less to you. For instance, while not everyone may be familiar with the definitions of open technology, I'm certain that you've heard of Linux. And of course, some may have heard of Open Source and Free Software, but of those, how many know, if there is one, the difference between the two.
- NT or Linux ?
- Patented or open technology ?
- Open Source or Free Software ?
All these dilemmas have something in common: freedom. And of course, lack of freedom, as found in the software industry today. Each will be developed to show how they actually point to the same concepts, how they came to be, and what we might be heading for tommorrow.
2. Context"Statistics show it: Microsoft Windows NT is the most popular operating sys tem, and is dominant in the server world."
Stop a moment to think about these words, because they are false, and I'll explain why, but they also play with reality. If someone does tell you this, ask them: for what class of systems, for what applications, and for which type of service ?
I can hardly imagine that the answer will be "NT is the most popular main frame operating system, and is dominant in the data warehousing world." -- you know, and I know, that this is simply impossible .
The problem here is _context_. What we're missing is a referential to place the statement in. Otherwise, like most statistics, it is meaning less . But why does it work so well ? Because of the iceberg principle: what's visible is what gets the attention, not what's under the water, even though most of it lies there. In the software industry this translates to the observation that what gets the most computer press coverage is not what's making the most money, but rather the technology that is most exposed to the public (or the other way around).
In other words: PCs sell, mainframes don't. It's much more likely that your boss has a PII-350 PC on his desk than an Amdahl 390. So of course he'll relate much more easily to the technology reading Byte or Computer World, which are very much PC centric and tend to oversimplify.
And when you hear "NT holds 60% of the market" -- is it:
You can throw these numbers in with the "market share" claims of Inter net Explorer and IIS.
- for the number of licenses sold last year ?
- for how many users per server ?
Now, we started out talking about freedom, and we've come to statistics and press tricks. You might think they have nothing in common, but they do, and it's all about one thing: monopoly and proprietary solutions.
Keep this in mind as the first piece of three in a puzzle to hold a market under control.
3. Intellectual property or poverty ? 1980-1990
Whatever the point of view, one thing is certain: patents are hard to claim, and even harder to defend. Ask IBM, they've tried. While at first they didn't believe in the PC when most other microcomputers had such a large advance, they changed their attitude when they saw how popular the PC was becoming, and decided to patent the BIOS and the design of the PS/2 computer. The same can be said of Apple -- the Apple II was a rather open computer, but they learned the lesson, and found a much simpler solution than IBM when the Macintosh came out in 1984: don't protect the technology, hide it . Make it inaccessible, or with great difficulty, and only to a selected few. If pos sible, make them pay for it at the same time -- after all, they're all beg ging for the privilege to develop for the trendiest and most stylish com puter corporation. Apple pushed it so far that people -- and a faire number of developers -- got sick of it and went to the competition; as ugly as it was, it was still cheaper to buy, and hey, you could develop for them AND pay your rent. So it was the beginning of the Microsoft Era.
- "Patents are a Bad Thing."
- "Patents are the only way to garantee the protection of our work."
Of course, proprietary software and hardware didn't wait for Apple and Microsoft -- long before their time, IBM had learned the trick, but at the time, and for the amount of money involved, they usually shipped in the box an extra engineer who knew his way around things, and made sure you didn't connect DEC printers to IBM minis.
Microsoft learned a few things themselves, and it was that applications were important, not the operating system. And in fact they did very little effort on the OS side, but did produce some rather nice applications in their time .
They succeeded in grabbing the PC server share of the cake in the end: devel opers who had seen nothing but DOS were in awe and users, convinced by the greatness of the Windows 3 (!) User Interface, in turn convinced management that Windows was a Good Thing. Management then heard of Windows NT 3.1, and seeing that the UI was the same, decided that it was a Good Thing, and in turn decided to forcefully convince IT departments all over the world.
There were some users who did complain about stability, and a few devel opers who missed the simplicity of the Macintosh, but overall the costs seemed lower, and Microsoft promised greater stability and a new, programmer- friendly API for Real Soon Now.
The rest went quite fast: OS stability barely improved, but users got... used to it, the Macintosh was slowly abandoned by Microsoft software, and when Win32  appeared, it broke compatibility with many programs and forced many software developers to buy a new license (Win16 had been too much exposed), or simply give up because it was too complicated to keep up with the changes.
As a side effect, Microsoft's own applications gained in speed using API calls that no one could find in the documentation, companies that made 3rd party Win32 development tools suddenly had their source licenses revoked , and prices on Microsoft applications went up again as they gained market -- since no one could make programs to match the performance or integra tion level of Microsoft's own, at least on Windows.
On another level, Windows piracy increased, but Microsoft left that alone -- in the meantime they were selling books to people who hadn't bought the software -- until they reached 90% of the market, both official and black, at which point they started to turn against the pirates .
Users were hooked, and those developers that wanted to go back to the Mac couldn't, because they had invested so much time in Windows software, and there was no Mac market to speak of anymore.
We started out to talk about freedom and on the way visited manipulation, con templated cheap marketing, and arrived at our last stop.
Hold on to the second piece as we try to find how the third fits in.
4. (Gratis ? Free : Open)Interestingly enough, there are other worlds than that of the Mac and the PC, and one you've surely heard of is UNIX in general, and Linux in particu lar. Linux has received a lot of press coverage recently -- IBM, HP, Oracle, Sun -- almost every "visible" player in the software and hardware world has praised Linux in one way or another and hurried to port their software or support Linux .
Why would they want to do this, and why now ?
The companies mentioned above are not philantropic, they're just out to make money like everyone else: they're not supporting Linux because they think it's a worthy cause, or because they think Linux and Free Software in general are better -- they just happened to pick the one that had the edge, and isn't likely to turn into another Redmond Giant  -- there will be no risk of los ing control of applications or simply seeing them get sunk by Microsoft-equiv alent products -- can you say WordPerfect and Lotus 1-2-3 ?
As to why they've waited today, there are two reasons: Microsoft cur rently has its hands tied by the U.S. Department of Justice trial, so it's unlikely they'll show their monopolistic attitude by punishing the other soft ware companies for doing this. And before 1992, there weren't any Operating Systems with no strings attached to them, usable on a PC. The fact that UNIX, through the combined popularity of Linux, FreeBSD and Free Software had been given a second chance , made a perfect occasion.
There a are other companies in the race than the big players we men tioned above. Some of them have heard of Free Software, and are trying to benefit from the higher quality development model it offers. Others have heard of something called Open Source, and thought it would be enough to give the source code, and sell the rest to become popular . Microsoft too has heard of Free Software, and they do not like it very much -- the furthest they have gone is freeware, or pusherware, like Microsoft Explorer: get the first dose for free, buy the next ones. Microsoft even admitted they were worried in a series of 3 internal documents that made their way to the press: the Halloween documents .
In them they explicitly admitted that Linux was becoming important, and that since Free Software was based on open standards, the best way to push them out would be to change those standards -- why attack when you can poison the water ?
It is in their advantage to play on confusion and try to convince users that their best interest lies in Windows -- scare them away from Free Software, take control of standards and protocols so that only Microsoft products can interpret them. Some of this has already happened, and it might happen some more.
5. Other horizonsYou now have the three pieces -- a market monopoly takes effort to build and afterwards hold, an effort in three directions:
So how do we now we are in a situation of software monopoly ? In most other industries, when one is dissatisfied with a product, one just goes out and buys from another manufacturer. You don't like Sony ? Buy Philips. Think Saab is too snob ? Buy Volvo.
- propaganda and manipulation
- predatory strategies and aggressive marketing
- corruption of standards
We could say the same thing in the PC software world, and tell people that if they're not satisfied, they can very well go out and buy another office suite or another operating system... except they don't. For one thing, it probably wouldn't work as well if there happened to be a Microsoft equiv alent -- who's never heard of the dreaded Word 97 compatibility problems ? -- in the case of software, or it simply won't run Windows software, in the case of the operating system. So you could believe Microsoft's bedtime sto ries, that everybody is satisfied with Windows. And accept it.
You could compare that to being told that you can only listen to a 1/3 of the FM radio stations, and listen to 1/10th of CDs if you used Sony instead of Philips, or that your Volvo can not drive on all roads because of incom patibilities between the tires and the asphalt.
Of course no one would accept this.
And when the computer press tells you that it's not that bad, since so many people are satisfied, then /why is it that so many people complain against Microsoft and telecom companies in general ?/
Because they _are_ monopolies. And when Microsoft's time had come and passed, will there, and should there be another one ?
- Even if NT ever ran on mainframe class machines.
-  «Studies show that 97% of all statistics are false»
-  It worked so well that they managed to misplace 35% of the the market share at the same time, and they still haven't found it.
-  Or did they ? They wrote the original Word for Mac, Excel, Encarta, but they did not write Word for Windows, FrontPage, or Microsoft SQL Server.
-  Which one ? 9x or NT ?
-  Citrix, Bristol Software
-  In some parts of the world, like South America, Microsoft knowledge ably let piracy develop until their software reached saturation, and then proceeded to threaten to sue entire governmental organizations if they didn't pay the licensing fees requested
- [8, 9] "Linux's success story is remarkable, but nothing is garanteed"
"Linux and the monopoly game"
Both articles by Nicholas Petreley who wrote the excellent series "The Next 10 minutes" on Windows NT:
-  "Pulling on the end of the rope"
-  "Why Free Software is better than Open Source" http://www.fsf.org/philosophy/free-software-for-freedom.html
-  http://www.opensource.org/halloween.html