The uncertain future of Windows NT

Recent industry comments on the strength of Linux are too meaningful to ignore

Original link = http://www.linuxworld.com/linuxworld/lw-1998-11/lw-11-penguin.html

Summary
An unabashed FUD piece on the future of Windows NT. One by one, LinuxWorld Editorial Director Nick Petreley examines the various means by which Microsoft may combat the threat of Linux and concludes that the future may be dim indeed for Redmond's premiere operating system. (3,100 words)
By Nicholas Petreley
When we laid the foundation for LinuxWorld, I tried to make it clear to my associates at Web Publishing Inc. that I didn't want the magazine to become an anti-Windows Linux-advocacy rag. I wanted LinuxWorld to be as courageous about being critical of Linux and Linux applications as it would be of other platforms and applications.

However, this month I'm afraid I'm going to have to break my own rule. In fact, I fully intend to wax rhapsodic over the advantages of Linux and spread fear, uncertainty, and doubt (FUD) about the future of Windows NT.

I find it necessary to make this exception for three reasons. First, a significant number of readers have indicated they're looking for credible pro-Linux ammunition in their effort to sell Linux to their bosses. Second, it's entirely possible that some of those bosses read LinuxWorld, and if so I can save the aforementioned readers a little effort by speaking directly to those they want to influence. And last but not least, I'm doing this because it will make this month's column ridiculously easy to write. Some of the best writing, in fact, has already been done for me. All I have to do is quote it -- which I shall do presently.

First, I'd like to warn Linux aficionados that you will probably recognize the following quotes. If your boss is reading over your shoulder, I ask that you to please refrain from giggling so you won't spoil the punch line.

As for those of you who don't recognize the quotes that follow, trust me. These comments are certainly authoritative. They are backed by hard research done by those whose job it is to understand Linux, Windows NT (aka Windows 2000, aka the W2K problem), and how the two compete. In case these arguments aren't convincing enough, I've included personal comments from a programmer and supporting comments from the president of a major corporation.

The Linux advocacy
We'll begin with selected quotes that testify to the quality and credibility of Linux and open source software.

Linux represents a best-of-breed Unix, that is trusted in mission critical applications, and -- due to it's open source code -- has a long term credibility which exceeds many other competitive OS's.

Linux is on track to eventually own the x86 Unix market and has been the only Unix version to gain net Server OS market share in recent years.

[Attractive Linux features:]

The GPL [GNU General Public License] and its aversion to code forking reassures customers that they aren't riding an evolutionary "dead-end" by subscribing to a particular commercial version of Linux.

When overall satisfaction with the OS's was calculated, Linux came out in first place. Linux was rated #1 in 7 of 9 categories in the DataPro study... [In order of preference, the study lists Linux, Solaris, HP-UX, NetWare, Windows NT, Digital Unix, IBM AIX, SCO UnixWare, and SCO Open Server.]

Using today's server requirements, Linux is a credible alternative to commercial developed servers in many, high volume applications.

And a personal testimonial regarding collaboration on Linux open source code:

I'm a poorly skilled Unix programmer but it was immediately obvious to me how to incrementally extend the DHCP client code (the feeling was exhilarating and addictive).

The anti-Windows FUD

Linux's (real and perceived) virtues over Windows NT:

I previously had IE4/NT4 on the same box and by comparison the combination of Linux/Navigator ran at least 30-40% faster when rendering simple HTML + graphics.

[According to the president of a major corporation, Windows NT 5.0] can stand more improvement in three main areas: configuration management, availability and reliability, and scalability. ... [This president] cited one Microsoft customer [using Windows NT] that automatically re-boots a particular server every four weeks because otherwise, the server will begin to bog down. Yet neither it nor Microsoft has been able to figure out what exactly is causing the problem.

Scalability, Interoperability, Availability, and Management (SIAM) are the most often cited reasons for using Unix over NT in mission critical, high-end applications. In today's Linux deployments however, scalability is not the driver as much as Interop, Reliabiliity, and Headless Management.

And a personal testimonial to how much more efficiently Linux is developed and debugged, as compared to Windows NT:

... due directly to GPL + having the full development environment in front of me, I was in a position where I could write up my changes and email them out within a couple of hours (in contrast to how things like this would get done in NT). Engaging in that process would have prepared me for a larger, more ambitious Linux project in the future.

The punch line
The punch line is simple. All of the above material originated at Microsoft. The president is Steve Ballmer, the major company is Microsoft, and the quotes are from an October 22, 1998 Network World Fusion article.

The rest of the quotes above have their origin in the now infamous Halloween documents, so named because the first one was leaked to open source guru Eric Raymond on or about Halloween of this year. (Use the following links to view the Halloween I and Halloween II documents in their entirety, along with Eric's comments.)

It is quite an understatement to say that the Halloween documents have caused a stir in the press. Reporters rushed to cover the story. Pundits voiced conflicting opinions about their origin. That dispute was quickly settled when Microsoft confirmed that the documents were genuine but were designed to stimulate discussion, not state official Microsoft policy.

The plot thickens
The fact that Microsoft admitted these documents were genuine elicited more controversy. Pundits had a field day trying to figure out why Microsoft so quickly owned up to creating the documents. Some suggested that Microsoft deliberately leaked them to fulfill an ulterior motive. Not everyone agreed on why Microsoft would do such a thing, but the most popular opinion seemed to be that Microsoft wanted to create the illusion that it does not have a monopoly in operating systems.

I find that interpretation rather amusing, even if does turn out to be true. In the first place, Microsoft doesn't need to pretend Linux is fierce competition at the server. It really is. And Microsoft's monopoly is so far confined to the desktop, not the server. That makes it even more unlikely that Microsoft was trying to pretend it doesn't have a monopoly, since it's the server market that is addressed in these documents.

Microsoft doesn't need to
pretend Linux is fierce
competition at the server.
It really is.

No, if these documents were leaked deliberately as part of a conspiracy, I can only conclude that Microsoft is engaged in a conspiracy to destroy its chances of dominating the server market. Rarely have I seen such powerful arguments to use Linux in place of Windows at the server.

Indeed, if the above quotes had appeared in the context of page-one news stories in the major computer trade publications (InfoWorld, ComputerWorld, PC Week, Information Week, etc.), one could only conclude the trades had already decided Linux is the future. And once that happens, the rest is usually a self-fulfilling prophecy.

The coming onslaught
Most mystifying of all is the fact that pundits tended to focus on the portions of these documents that describe how Microsoft might fight Linux and open source software. By doing so, pundits awarded Microsoft what is likely to be its only victory.

As Mark Stephens (aka the PBS version of Robert X. Cringley) observed in the November 5 installment of his column, "The Pulpit," by focusing on the fear that Microsoft might beat Linux and open source, open source advocates have allowed Microsoft to put them on the defensive. To quote Mark:

What a week ago was a discussion about the inroads Linux and Apache have made against commercial software has suddenly and instantly been redefined into a discussion of the threat Microsoft poses to Linux and Apache, and what those two development efforts have to do to survive.

This is especially ironic because only a few of Microsoft's tactics have even an ant's chance at a dance. My interpretation of these tactics are as follows:

Beyond the first tactic, Microsoft has little hope. Even Microsoft admits that the second tactic (FUD) is unlikely to work. To quote from Halloween I:

The GPL and its aversion to code forking reassures customers that they aren't riding an evolutionary "dead-end" by subscribing to a particular commercial version of Linux. The "evolutionary dead-end" is the core of the software FUD argument.

Microsoft can attempt to pollute open protocols with proprietary extensions. But it is unlikely to do so with any success, at least when it comes to Internet and intranet projects. Bob Denny eloquently explains why in his Chicago Tribune Silicon Prairie article, "Buying into Microsoft."

Denny describes a scenario in which you, the IS decision maker, build a cool intranet application based on Microsoft's flavor of dynamic HTML, XML, Microsoft XMS extensions, and ActiveX, then one day your boss asks you to open up the system to the company's customers and partners. He elaborates on the consequences:

Oh no! Your customers [and vendors with whom you intend to do business] have various browsers out there and some of them won't display the pages your systems team developed. You issue orders to develop browser agnostic Web pages. And now those pesky IETF and W3C standards get in the way! Your systems people start gasping for air because it means the loss of some coolness, and worse, some drudge-type work. Or maybe you decide to develop a parallel set of pages for outsiders and maintain both. And who's gonna pay for this?

Live free or die
The fact is, unless Microsoft can take over the Internet server space overnight (thus eliminating the need for open protocols in order to grant every possible client seamless access to Web-based data), it is nearly impossible for Microsoft to pollute open protocols with proprietary extensions. Most developers (at least those with a trace of intelligence) will avoid them in order to avoid the possibility they would exclude even a single potential customer.

Unless Microsoft can take
over the Internet server
space overnight, it is
nearly impossible for
Microsoft to pollute open
protocols with proprietary
extensions.

The same principle holds true for the fourth item -- polluting the market with complex and proprietary middleware. Doesn't Microsoft wonder why CORBA is light years ahead of DCOM+ in terms of both technology and market acceptance? How about the reason why Java has been so wildly successful in spite of delays and blunders by Sun? The answer is simple: CORBA and Java (and the like) don't corner the user into a single vendor or platform.

The possibility that Microsoft could actually hold up Linux development through patent and copyright lawsuits is patently ludicrous. What could possibly be more damaging to Microsoft than if it actually won such a lawsuit? With the possible exception of BoB, everything Microsoft has produced is a knock-off of someone else's innovation. If Microsoft won its battle, it would surely lose the war. Companies would tie up Microsoft for decades with copyright lawsuits based on its own precedent.

And think of the possibilities of a countersuit. What would happen if Linux representatives subpoenaed Microsoft source code and discovered it contained code licensed under the GNU General Public License? Microsoft would be legally required to release its source code to the public. (This is pure speculation on my part, by the way. I have no evidence that GPLd code exists in Windows.)

And by the way, who exactly would Microsoft tie up in litigation? Linus Torvalds? Now there's a way to bolster your corporate image.

Draining the pool
That leaves two strategies: thin the OSS/Linux talent pool by hiring away the best and brightest contributors to Linux, or woo away Linux application developers by providing inexpensive or free entry-level tools.

Microsoft has already attempted to hire away Linus Torvalds's right-hand man, Alan Cox. Alan didn't bite. (Alan maintains the 2.0.x production versions of Linux kernels.)

I predict that some Linux developers may bite, but very few. Many of the people who contribute to Linux do so because they want to provide a great alternative to Microsoft products. To join Microsoft would be an act of treason.

In addition to the religious issues, the appeal of working for Microsoft is waning. The primary draw Microsoft has is its promise to make its employees millionaires as the company stock price soars. Salary-wise, Microsoft programmers are said to make about 75 percent of the going rate for programmers at other companies.

This is a critical time for Microsoft and its stock. It is unlikely that Microsoft will be able to build much momentum for Windows NT until well after the year 2000, since few customers will initiate a massive migration to NT in the middle of a Y2K crisis. As the public continues to lose confidence in Windows NT and especially if Microsoft continues to falter on other fronts (note that interoperability problems between its own products have recently surfaced, service packs have been recalled, etc.) the stock growth for Microsoft will continue to slow. If the stock price levels off or drops it won't look quite so appealing to work for a company at 75 percent of the pay you'd get elsewhere.

But let's suppose Microsoft is willing to compensate for that. Suppose it shoots the moon and offers a $2 million sign-on bonus to each and every one of the 100 best contributors to Linux. And why not? $200 million is petty cash to Microsoft right now.

If so, the company succeeds in alienating the talented programmers it currently has on staff who didn't get $2 million bonuses, while creating a temporary hole in the Linux community. And rest assured, the hole would be temporary. The reader mail I get as a result of my InfoWorld column testifies to the fact that universities are churning out Linux experts in droves.

Woo woo
The Microsoft documents also demonstrate a total lack of understanding regarding Linux developers when it suggests it might woo away developers with inexpensive entry-level tools and limited open source code. In the first place, Linux development tools are quite satisfactory, thank you very much. Heck, Microsoft could give away all its tools. Ask any Linux developer if he or she enjoys using gcc, Tcl/Tk, Perl, Java, Python, Gtk, etc. Then ask them if they would prefer to use Microsoft C++, MFC, Access, or Visual Basic. Anyone care to place a bet on the outcome?

There is also more to ease of development than a drag and drop visual development environment. There's the architecture of the system itself. The Halloween documents speak to the difference between Linux and Windows NT in this category:

An important attribute to note which has led to volume drivers is the ease with which you can write drivers for linux, and the relatively powerful debugging infrastructure that linux has. Finding and installing the DDK, and trying to hook up the kernel debugger and do any sort of interaction with user-mode without tearing the NT system to bits is much more challenging than writing the simple device-drivers for linux. Any idiot could write a driver in 2 days with a book like "Linux Device Drivers" -- there is no such thing as a 2-day device-driver for NT.

Finally, Microsoft suggests it may win over Linux developers by opening up portions of Windows NT source code.

Again, this demonstrates a total lack of understanding of the philosophy that drives Linux. The single most reassuring aspect of developing for and adopting Linux is the knowledge that the GPL is an all-or-nothing proposition. Linux is a safe choice because there is no possibility we'll wake up one day and discover the next version of Linux is closed and proprietary.

Microsoft endorses Linux
How does this all figure into the irrational way so many people make decisions? Whether or not anyone with technical knowledge wants to admit it, many purchase decisions are made based on public perception over quality.

If you read my November 2, 1998 InfoWorld column "Down to the Wire," you're already familiar with the factors that go into product selection when it comes to Dilbertian pointy-haired bosses (PHBs). To paraphrase my original formula, the perceived quality of a product (Q) is determined by the number of positive comments (P) divided by three times the number of negative comments (N) divided by the article's distance from Page 1 (D). The formula is therefore Q = (P/(3*N))/D.

Until recently, Linux suffered from poor public perception. It was regarded as a hacker OS with little or no qualified technical support. Most trade publications had already decided Linux was irrelevant, and that Windows NT would overwhelm even the most well-established versions of Unix.

One cannot deny that the
Halloween documents contain
some of the most convincing
arguments for using Linux
instead of Windows NT.

This thought rankled Linux fans, because they knew Windows NT was a colossal market flop by any rational standard. The technology in NT is still years behind what Microsoft promised it would deliver in 1994, let alone how it compares to current versions of Unix. (Is it any wonder Microsoft had to abandon the code name Cairo? Windows NT 3.5, 3.51, and 4.0 couldn't begin to live up to the hype associated with that name, and Microsoft isn't even promising all of Cairo's features for Windows 2000.)

Nevertheless, Microsoft, along with a willing computer trade press, created the perception of unstoppable momentum for Windows NT. They did so by endlessly repeating unchallenged quotes from Microsoft project managers, sales representatives, and research analysts who, although perhaps intelligent, well educated, and well meaning, probably couldn't compile a to-do list let alone a "Hello world!" program.

Then, as Linux momentum in the press suddenly surfaces almost out of thin air, the Halloween documents appear. And while they set some Linux defenders into a frenzy, one cannot deny that these documents contain some of the most convincing arguments for using Linux instead of Windows NT. Arguments even a PHB can appreciate.

So how does this all figure into the irrational way so many people make decisions? It's difficult to predict. It will be the ultimate irony if these documents prove to be the pivotal catalyst for the demise of Windows NT and the success of Linux. Only time will tell.

Discuss this article in the LinuxWorld forums
(You must be a subscriber to particpate. Read our forums FAQ to learn more.)

About the author
Nicholas Petreley is editorial director of LinuxWorld, columnist for InfoWorld and NT World Japan, and an associate producer for IDG Conference Management.