Before leaving for the CRM and Support Services 2000 Conference and Expo in DC, I perused the schedule for must-see presentations. In the midst of the experts, consultants, and geeks was one of my favorite speakers, Mark Minasi. Mark is usually pretty popular among conference-goers, and it’s no wonder. I’ve heard him speak a few times and always found him to be intelligent, informative, and very funny. Intelligence and information are great, but after a few days of information overload, funny is greatly appreciated.
Mr. Minasi is not just a speaker. He is the best-selling author of Mastering Windows NT Server and, more recently, Mastering Windows 2000 Server . He is a regular columnist for Windows 2000 Magazine (formally Windows NTMagazine), and Chief Scientist at Minasi R&D. As Mark is a published writer, I thought he’d take pity on me and agree to be interviewed. I was right. We scheduled a breakfast appointment for the following morning, and here’s what Mark had to say.
TR: Chief Scientist. That’s an interesting title. How did it come about?
Mark: I used to have a real company, 25 employees. We pulled in several millions of dollars a year, went around the country and did consulting and educating and things like that. …We don’t say training. That’s a bad word… And I got tired of it. But in the process, I went to a lot of meetings for training associations. They would say, “Okay, everybody please go around and say who you are.” I remember hearing—I’m going to have to change the name here—a man standing up and saying, “Hi. I’m Mel Jones of Mel Jones and Associates.” This may seem like a simple thing but if it’s just you, say it’s just you. What’s the associates? The answering machine? The dog?
Being president of a one-person company seems Viagra-challenged to me. So I thought, if I’m going to have a title, President is not all that exciting. President sounds like a guy that wears suits far too often and doesn’t actually do anything. I guess he’s the point man. Contrast Chief Scientist. What better title could you imagine?
TR: I don’t really like the titles they give computer support people today. Systems Integrator is a common one, also Client Support Analyst.
Mark: Well, support is a terrible phrase. If you want support, get a jock strap.
TR: Do you have any recommendations on what we could start calling ourselves?
Mark: Handy People or IT Fix-it Folks. Actually, Fix-IT Folks. That’s a little too precious. We all want titles that sound like there was a good reason for us to get that social science degree. None of us in the business are actually using that degree in sociology or economics or theatre that we got.
TR: Let’s change the subject to something a bit more technical. You speak and write a great deal about Linux and Windows NT and Windows 2000. Which do you prefer to use?
Mark: On the desktop there’s no question, 2000. You don’t use Linux on the desktop. Maybe you will in a couple of years. Linux is a society of hackers building cool stuff to impress other hackers. Which means you get great server products, great mail servers, great Web servers, pretty good Web browsers. Things like GUIs, no. “So we’re going to do GUIs? Why? To make it easier for those users, those newbies to come in? Naaa.” I guess it’s a nice thing, but it doesn’t get you in the Warrior’s Hall of Fame. Which maybe means there’s a commercial opportunity, I don’t know.
Right now hackers are building a GUI called GNOME, and it’s terrible. It’s big. It’s slow. It’s fat. It’s like they’re trying to make Linux as inefficient as Windows.
So, what do I use on my servers? I’ve done most of my work on NT. So, for a good long time my backbone servers have been NT 4. My main server, however, acquired a hard drive problem, just in the middle of a crisis. …Computers have urgency detectors on them….
So I moved over to Windows 2000. My domain controller is currently a Windows 2000 Domain controller. But the mail server is a Linux box because my e-mail server, believe it or not, will not run on a 2000 server. I had this commercial mail package running on NT 4, no problem at all. I didn’t even think twice. I thought I’d build a 2000 server, but I put this product on and it didn’t work! I called the vendor and they said, “No. It’s not going to. For $650 we can fix that.” I thought, well, I could use this Linux box, which is free and it does a very good job. So, I’m an integrated, open kind of guy. I’m an enterprise kind of guy.
TR: Of the different systems you use, which do you find easier to support?
Mark: I’m not a Linux expert yet. I’m becoming one. Given the choice of supporting Windows or supporting NT—that’s an easy one. I would support NT. It’s more reliable. It’s hard to get hardware to work on NT. It’s easy to get software to work and it’s easy to keep software working. And, if something goes wrong you’ve got really good diagnostic tools. It’s actually possible to smell a memory leak. It’s possible to discover bad drivers. With Windows 2000, even easier.
Windows 2000 has this very cool thing everyone needs to know about called the Driver Verifier. What makes systems crash is that drivers go places they are not supposed to go. They reach outside of their memory and modify some area. Driver Verifier wraps the drivers in security blankets, so to speak. If they try to reach outside, the sensors detect this. The system blue screens, and it points the finger and says, “That’s the bad driver.” Imagine how cool this is. This is the situation we’re in now with NT 4. You kind of think it’s this driver, but the vendor is doing this. [Mark points fingers in opposing directions] You turn on the Driver Verifier, and it says, “I’m, sorry. It’s YOUR driver. YOU wrote it. IT blew up. FIX IT!”
TR: The system blue screens, and you can tell from the blue screen what’s wrong with it?
Mark: It’s straightforward. Actually, reading blue screens is very straightforward. Typically, on the blue screen, up top you get four very long numbers. In one of the most common blue screens, the third number is the address where the bad thing happened. Below that is a list of nearby drivers and their exact addresses. So really, all you’ve got to do is pull that address off. All the rest are ones and zeros. It’s the only one that looks like a number. There’s an excellent book on NT that explains this.
TR: Is that Mastering Windows NT Server by Mark Minasi?
Mark: As a matter of fact.
TR: What changes do you least like in 2000?
Mark: Whenever Microsoft decides to take a step, I think they take too big a step. Let me roll back and talk about OS/2. OS/2 is an operating system that came out in December 1987. What problems were we trying to solve? 640K sucks, GUIs would be nice. We’d like to be able to run more than one program at the same time. But let’s peel back. What was the real problem? Memory. Whoop! Whoop! Whoop! Memory was the number one problem.
Microsoft could have easily, in ‘84, come out with a DOS 4, which just had a bigger memory model. That’s all. So let’s just kick that baby out. Then it becomes possible for vendors to make relatively small changes so you can run the old programs as well as the new memory-enhanced programs and, golly, that would be easy. And then, six months later, let’s add multi-tasking. NOT graphical multi-tasking, but like Linux does it. [Alt] [F1] to see the first box, [Alt] [F2] to see the second box, and so on. That way I can run a couple of programs. No. we can’t cut and paste, but maybe we can add that a little later, because when we go to graphics scary stuff happens.
One of the things DOS could do, DOS could run if it was on its knees. You didn’t have to have all of DOS running. Windows can kind of run on its knees as well. You can run in Windows safe mode. You can actually boot Windows to a safe mode command prompt, which means it’s broken but you can get it up. In contrast, the terrible problem with NT is that it won’t run on its knees. It’s ridiculous that I have a server that will blow up because of the graphics drivers. That makes no sense at all. You’ve got no business putting a screen on a server anyway. You should be remotely controlling it over a Web-based interface or something like that. We should have headless machines.
What Microsoft should have done was create big memory first, then multi-tasking, and then graphics as an add-on. This is kind of what they eventually did with Windows—not a bad thing. Then there’s Windows 2000. Windows 2000 represents paradigm shifts by the wheelbarrow. I think I would have done it a little more gradually. Active Directory is a big step. I would have done a release based on Active Directory, but I would have made it something you could bolt onto NT 4 real easily. They didn’t do that. So they took big steps, but consider that first step. It’s a lulu.
TR: You write for Windows NT Magazine, now known as Windows 2000 Magazine. You often get laughs at Microsoft’s expense. Does that ever make your editors uncomfortable?
Mark: Mark Smith is our Group Publisher and just general Grand Poobah. Mark is an interesting character. I think the reason the magazine is so successful is that Mark is not an editor. Mark was the network administrator at Duke Publishing, a big System 3X shop. If you did AS400, you would know their name.
Well, NT comes around, and Mark, who is their Netware administrator, actually, comes to the publisher, David Duke—not that one—and says, “We have to do a magazine on this, and I want to be the publisher and editor in chief.” Mark is an amazing guy. Quite honestly, most editors I’ve worked for know a lot less about the subject than I do. But Mark—his finger is on the pulse, and he’s been in the trenches. His hands are very dirty with networking. This is why, I think, if you pick that magazine up you will see that there are always three or four articles, if you are a network admin, that you want to read.
I’m like everybody else. I get 5,000 magazines a month, none of which I look at. There are probably four magazines I always make time for, Sky and Telescope, Discover Magazine, Star Date, which is another astronomy magazine, and Windows NT Magazine. I don’t even get PC Magazine anymore. You know, if I want ads, I’ll just get on the Internet.
TR: Now that Mr. Gates has returned to development, what would you recommend he work on?
Mark: I think that software is like any other technology business in that you start out and in the early phases and it’s, “Oh my god, we can do this? That’s so cool.” The people who buy 1.0 stuff are the people like me, the early adopters. But then we start evangelizing, and then 2.0 comes out and 2.0 is always, “Now that we’ve thought about it we really need some of this other stuff.” Plus, they fix some egregious bugs from the first one. Version 3.0 though, that’s when the technologists start to take a back seat and the marketers start coming out. By the time you get to version 5 or 2000 or something like that you kind of run out of juice.
Look at Windows. Windows 3.0 came out, some big changes. 3.1, bug fixes. Windows for Work Groups comes out, 32-bit stuff. It has a 32-bit interface. It’s 95, really. Windows for Workgroups was pretty close. Now, what’s Windows 98? It’s Windows 95 1.1. You can see that they are running out of juice. They need some new ideas. What’s really left? What’s really left? I mean, other than being able to get a jack installed in the base of my skull and just NCLK, you know, just jack in and say, “You know what I want. Just do it.” Beyond that, I guess talking to it in English. That’s decades away. Talking to it is years away. Wearable computers? I’ve got a Palm Pilot. I guess that’s wearable. I think at this point, what’s left?
Automobile business, same problem. In 1922 the question really was, “What interesting things can we do with a car? I know. Lets put lights on it so you can drive at night.” Cool idea! “Let’s have a windshield so you can drive more than 20 miles an hour without goggles.” Hey, great idea! Killer apps. Killer technologies.
By the time you get to the 1957 T-bird, you know, there’s not that much difference in the 1957 T-bird and the 2000 Camry. When I say that what I mean is that there are power brakes, power steering. You name it. All that stuff is still there. They ran out of features. So the auto industry began to look for ways to keep people buying. And the industry, quite deliberately, radically redesigned the look of the car every three years. So, if you had a four-year-old car, it looked like a four-year-old car, fins, stuff like that.
You know the end of this story. They missed out on the big thing. If in 1967 you wanted a small, good-on-gas-mileage, reliable car for driving to your job six miles away, you couldn’t buy one. Japan swooped in. The Japanese didn’t take the market away from us; we gave it to them. In the same way, the software business, I think, has moved to the point where the best feature they could offer would be quality. I mean I’d like to see the feature war end and the bug war begin.
More from Mark Minasi on Bill Gates
I would love to see a magazine ad, a print ad that says “An independent software technical laboratory demonstrated that Word has 41 percent fewer bugs than its closest competitor, WordPerfect. Buy Word.” I’d love to see an ad like that. But I don’t know if it’s ever going to happen. Actually it will happen. The question is: What country will the vendors come from?
I mean it’s not Japan. It will be India and Germany. They are producing some extremely high-quality stuff. For example, I like to tell the story that Windows 3 was a 2-MB piece of code with 5,000 known defects. I’ve gotten on the radio and people call in and say, “You’re crazy. You can’t have low-bug software. It’s not possible unless you go for the space shuttle at billions of dollars.” I tell them the Windows 3 story and I say, “Actually I have a microcomputer I carry with me all the time that has 2 MB of code and two known defects, the Motorola StarTac. And by the way, the code was written by a shop in Banglor, India. There are processes; there are techniques.
TR: My last question was going to be what you think the greatest challenge is for the industry today, but you just answered that.
Mark: Reliability. It’s reliability for a whole lot of reasons. We Americans own the software business in a way that we never owned any industry, not cars, not steel, not ships. God, I’d hate like hell for us to lose this. According to Malcolm and Ron, last year Americans spent 65 million minutes on hold, waiting for their Help Desks. Imagine if we had that time back to play with our kids or just put our feet up and have fun. Y2K bug. We spent 500 million dollars fixing that. Imagine if we spent that money improving the lot of the poor or just had a great party.
That segues me to another thing that irritates me: the media’s coverage of Y2K. I was interviewed probably a hundred times last December. In the beginning part of December the questions were, “Are the missiles going to fly? Do we need our bunkers? Are we going to die?” By the end of the month it was pretty clear things were going to be okay and I started hearing, “Hey, did those techies cry wolf? Is there really something there?” I was at the anchor desk of the local television news, the CBS affiliate in Virginia Beach.
The anchors were saying to me, “So, was this just a goof? Nothing really happened. It was much ado about nothing.” I said, “Well, wait a minute. We are an industry that’s known for blowing deadlines, and we got it done. We got it done on time and we made it look easy, and that’s the thanks we get.” There should have been a ticker tape parade down Fifth Avenue in New York for the Geeks. I mean, here’s a New Year’s that should have been a millennial event for many. Yet, many of the people who are reading this Web page probably got to spend New Year’s with a plastic cup of non-alcoholic champagne in one hand and a mouse in the other. And that was the thanks they got.
Actually, I have a proposal for that. As anyone who knows how to do math could tell you, the millennium doesn’t start until next year, not this year. I propose that we refer to the one that starts this year as the Broker’s Millennium and the one that starts next year as the Geek’s Millennium. We’ll have the Techie’s Millennium. Then we can have a proper celebration.
I was right. This interview with Mark Minasi was fun. If you enjoyed it, be sure to take a look at Mark’s latest book, The Software Conspiracy . I got it at the conference bookstore, and it’s quite eye-opening.
Pat Vickers is an MCSE currently with Sprint. To comment on this interview, please post your remarks below or follow this link to write to Pat .