'Threat' is probably the wrong word. Two things happen when a new product appears. It can either potentially take market share away from existing products, or it can expand the market as a whole. Imagine it was 1972 and we were talking about computers and IBM. In the end, IBM is making plenty more revenue today than 30 years ago. The same thing will happen with operating systems.
Microsoft has a big market share for several reasons. One reason has been the fact that you can buy one copy and put it on a bunch of machines. I'm not advocating piracy. I'm just stating a fact. There are people today who won't use XP because they want to put it on multiple machines, but that's the license.
Market realities forced every software vendor, around 1983, to stop copy protection. By 1988, there was not a major product that used copy protection. That had a lot to do with the growth of the industry, but it's all changed now. Microsoft has thrown the gauntlet down -- and they have the right to do this.
But I would say they only have the ability to do this because they are a monopoly. If there were a lot of competitors out there, or if one version became a pain to install, people [would] stop using it. We saw that with [IBM's] OS/2. People used Windows because it was easier to install.
But now Microsoft has made it harder to install their software, and people are going to begin to look around, and they will also say, 'the heck with it; I'm not going to upgrade.' It's one thing about Linux, you don't have to worry about license fees. But Linux won't be right for everyone.
I'm not a Linux expert, but five years ago I wanted to learn Linux. Let's be clear. The Linux GUI is a pathetic little toy compared with the Microsoft GUI. Linux, in its heart of hearts, has always been a command line-based OS. If it's your job to keep networks running, you've got to dig in and learn some of this command line stuff.
It's got some powerful appeal. First, command lines are parsimonious in terms of bandwidth. Sending a few characters costs me nothing. And there's scripting. Administrators do repetitive tasks. GUIs are great for something I do once a year and don't remember. But command lines -- most stuff should be done there, where they can be packaged and scripted.
So, is Linux a threat to Windows? Is roast beef a threat to a salad? They are different things. What about Linux benefits on the server?
On the server, Linux can be greatly simplified. One way to secure and speed your system is to take out the trash. I guarantee you have features you have never used. You can shut those things off, and people are slowly starting to do that. It's normal to take a Unix box and rip things out.
For some items, Linux is a good competitor to Windows 2003. My company may need a mail server, a DNS server, a router, a Web server. All of these things can be done with Linux, and the price is right. It's expensive to buy different servers in the Microsoft world, but not so in Linux.
The guys at Red Hat also did a smart thing. For $1,600, you get one year of any question answered by experts. In the Microsoft world, if you ask one question it costs $245. If you ask six questions a year at the Microsoft rate, you are ahead of the game. The biggest thing holding the Linux world back is that no one wants to pay for anything. Do you think Microsoft has recovered from the ill will caused by last year's licensing flap?
They haven't made any friends doing this, but I think they had to do it. With all software, the first few versions, you've just got to have. But by the fifth, ninth or 10th release, it may be good stuff, but it just doesn't grab you by the throat. Upgrading is a balancing act. You've got to buy things, get new hard drives, convert your documents, train everyone and find bugs. If new features are getting less interesting and your costs are growing, you start to say, 'I don't see why I should upgrade.'
Almost every upgrade I've seen has had positive value, but the marginal value is getting smaller. Who should upgrade to Windows Server 2003?
Anyone who is committed to Active Directory will find that Windows Server 2003 is the much-awaited 1.1 release. It fixes a lot, but not all, of the problems in Windows 2000. Also, security is easier. XP and Windows 2003 are more secure out of the box. What's your view on Microsoft's efforts to design a collaboration strategy?
Microsoft has enough money to throw around. They can buy this or buy that and see what they come up with. Sometimes these things become integral. Collaboration is a big problem and if they can solve it, it would be big stuff.
The big problem with our business is, we've been pre-sold and over-sold by science fiction. We have insane expectations that [have] been built by fantasy. We have cell phones modeled after the flip phones you saw on Star Trek. I'll bet those guys were never in cell hell.
We finesse the mundane realities in our modern fantasies. Data transfer time is never an issue. Storage is never an issue. Power is never an issue. We need better energy in our batteries. But we should have great expectations. Microsoft is working on it, and good for them. I hope they solve it. So people are changing their software-buying habits?
In 1999, things started to change. We had the Y2K lockdown. And then, after 2000, the world didn't end. Then there was the tech stock crash. By March 2001, people started looking around, and they saw that the money wasn't coming back. Everything slowed down. At this point, the CEO is saying, 'we haven't upgraded, and nothing bad has happened. The business is still running.'
Lots of businesses are still running Novell 3.1 and Windows 95. If your job is to make something non-technical, you can live pretty well and, in fact, that's what is happening. People are saying 'why should we upgrade?'
The fact that people are upgrading slowly means we will have a broader range of stuff to support. In two years, I don't think it will be unusual to see someone support Windows 95, 98, 98SE, ME, 2000, XP and XP professional. That's a lot to know. There has been a steep drop-off in students who attend Microsoft training classes. How will this play out in terms of a company's ability to keep up with the demands of new technology?
First, people don't go to training to keep up with the demands of new technology. Half of the people who go do so so they can take the certification test right away. Microsoft's curriculum -- the guys who put out the tests -- must do it quickly. They don't get a chance to revise it, and it leads to obvious problems.
People go to these classes to get certified. So why aren't people getting certified? There was a time when knowing NT could command a lot of dough. Having certification isn't cool anymore. Having an MCSE doesn't mean much.
FOR MORE INFORMATION:
Meet Minasi, hear his keynote address, and attend a panel discussion "Will Linux spell doom for Microsoft?" at Enterprise Windows Decisions, May 14-16 in Chicago. Conference attendees can also meet TechTarget editorial vice president Paul Gillin and Samba Team cofounder John Terpstra.
Want to get an advance copy of the first chapter of Mark's new book, Mastering Windows Server 2003? Update your member profile! Click here for details.