musings

by Rik Farrow
<rik@spirit.com>

Rik Farrow provides UNIX and Internet security consulting and training. He is the author of UNIX System Security and System Administrator's Guide to System V.

I am going to try something different for this column. I am going to pretend to be a journalist. I have no training in journalism per se, although I was given a two-day class in writing for magazines by UNIX World. For this performance, I will actually interview a source.

The person I have chosen to interview is Richard Diemer. Mr. Diemer is a tool-and-die tradesman who has worked for General Motors for 24 years. It was actually my wife who brought him to my attention, after he had shared some of his insights about working in a modern manufacturing plant and how that had affected his life.

Diemer works the swing shift, mid-afternoon to evening for those of you who have never experienced factory life. (I have.) His specialty, tool-and-die maker, was critical to the manufacturing of automobiles. Those graceful fenders, side panels, and hoods get stamped out of pieces of steel when the tool and die are precisely matched. While the design of the fenders and other parts are left to the designers and the engineers, it used to be that tradespersons like Diemer actually kept them working. A new (or worn) die might need a bit of tweaking, just a few millimeters shaved off, and the same few millimeters added to the tool so that the fender comes out perfect.

Today, Diemer watches the process from behind a window. The tools and dies are computer-designed, and he is rarely allowed to touch them. In the cause of efficiency, the smallest possible piece of steel is used for each part, instead of a slightly larger piece that might need some trimming afterward. In the past, if Diemer had noticed that the tool and die were not producing a complete fender (the steel extending to all the edges), he would have taken the stamping press offline while he and fellow tradespersons remachined the tool and die so that it worked correctly. But now he gets in trouble for stopping the process or even getting on the other side of the glass, up close to the computer-driven machinery.

Diemer has six years until he can retire. His job satisfaction is at an all-time low. Any needed machine maintenance is scheduled for breaks and lunchtime, cutting him off from camaraderie with his fellow workers. Once a master of his trade, he is reduced to tending the machines, and his suggestions, based upon over two decades of experience, are mostly ignored.

Sysadmin
You might wonder what this has to do with you. Programmers and system administrators, especially experienced ones, are at the top of the heap, able to demand great salaries and jump between jobs with the greatest of ease. Just like Diemer, who was also at the peak of his experience, until he was marginalized. If it weren't for the union, GM would have fired him, er, downsized him, long ago.

Max Southall <max@prninfo.com>, the MIS director of Kelme USA Inc., wrote a thoughtful letter that appears in this issue's letters-to-editors section. Here is a snippet of what he wrote to me:

Pretty well the only practical solution I've actually seen implemented by those who've tried to stick by Windows is the hiring of additional and progressively cheaper bodies to try to keep it all going somehow. And concomitantly, the laying off of the fewer more expensive bodies, because, as they say, it doesn't matter how smart you are under the MS scenario, because it takes just as long to reboot yet again. . . . In any case, costs keep rising and the level of service sinking. . . .

I just can't stand the thought of system administration being reduced to carrying a CD fanny pack from user machine to machine, forever. And that's what's happened to some of my formerly UNIX colleagues.

I am not simply writing another diatribe about Microsoft. Today, sysadmins carry critical information in their heads about the working of systems and networks. Their experience is great; their tools include Perl and shell scripts. Like Diemer, they are great and talented craftspersons, or they would not be able to do their jobs.

Yet the old way of fixing each problem as if it were unique is doomed. There are not enough talented people to support a world of computers and networks, regardless of whether they are running some version of Windows or of UNIX. The age of fighting fires by creating one-off solutions is passing, and so might the high times for sysadmins. We need solutions for managing large numbers of systems, adding users, installing software and patches, and changing configurations — and those tools must be used everywhere.

USENIX and LISA conferences include solutions to these problems in the papers track. But so far, none of these solutions has worked either well enough or universally enough to be widely accepted. The open-source movement has the potential for creating solutions that will make the future of system administration one of creativity and pleasure, not one of rebooting systems and wearing a fanny pack of CDs.

There will still be a place for shell scripts and Perl. No one system will fit all sizes, and there will be interesting and well-paying work for everyone who is capable. But not if we remain stuck, thinking that the world of the tool-and-die tradesperson will be with us forever. Just ask Richard Diemer.

Perhaps the tools that will save us already exist. A single tool that could securely and reliably distribute files and adjust configurations could do the trick. In the early '90s, I thought that perhaps Tivoli had the solution, but it was dreadfully slow, it was proprietary, and now it's also terribly expensive. I have seen several tools, written up in USENIX and LISA proceedings, that seem to come very close. They needed polishing to make them usable by anyone and portable to anything that runs an IP stack.

I am not the person who will choose and promote such a tool. I humbly suggest this as a path that should be taken, as it behooves us not to make ourselves obsolete, replaced by inferior technology, as in Southall's comments. What he wrote about is real. I had already interviewed Diemer before I received Southall's email. I have seen the same things coming to pass, and worry about the future of my friends, even the future of civilization.

I believe that there are broader issues here than just system administration. I mean more than just Visual Basic versus real programming. Are we going to design our own future, or just let it happen to us? If we do that, it will be designed to maximize profits, not to the advantage of everyone.

Management
Okay, I am now stepping off my soapbox. I received a book to review about management techniques that I'd like to share with you. In the world we currently inhabit, one of the ways that you make more money as you grow older is to slip into a management position. This is largely a function of human resources, the group that keeps charts describing exactly how much salary each named position may be granted, regardless of reality or market forces.

Management is never easy. We certainly weren't taught good techniques in school, and some of us chose to work with computers instead of people because it was easier. So when I started rereading this book and listening to the tape while I drove to the airport, I realized that this is the book that could help many people in their dealings with subordinates.

The book is by Adele Faber and Elaine Mazlish. You might not recognize the authors, as they won't appear in the Business Week top-sellers list. The big focus of the book is learning how to listen and then basing your responses on what you have heard. Listening is in itself a lost art. Most often, people will be planning a comeback rather than really hearing what the other person has to say.

The book is entitled How to Talk So Kids Will Listen and Listen So Kids Will Talk. No kidding. I used this book when learning how to be a parent. I have two grown stepchildren, and when I compare my relationships to my adult children to those of other parents, I know that I have been successful. Other parents have wondered at my success but would not take the time to learn some simple techniques.

One of the techniques involves acknowledging feelings. It is as simple as this. One of your workers comes to you, and he is obviously angry. Instead of explaining to him why you can't do anything about his problem, you say, "You're angry." By acknowledging the anger you get it out into the open; you permit the person his feelings, rather than denying them. As the anger subsides, you can begin to talk about possible solutions. When solutions are really impossible, the book has answers for that as well.

Of course, everything in this book is geared toward children. But we learned our own personal-relationship skills when we were children, and our parents were using the same techniques that were handed down to them by farmers and herders from thousands of years ago. Primitive techniques, in other words. We have not been proactive about social skills any more than we have taken an active interest in shaping our own future.

Communication is another of the problems that Diemer has. His bosses proclaim that they listen to what the line workers have to say, but in reality they don't listen. I had hoped that the reorganizations of the '90s were about streamlining management so that the head of a corporation would be closer to those actually doing the work. Boy, was I naive. Reorganization turned out to be lip service (are you listening, George Bush?), not a way of improving communication but, rather, one of increasing profits (and the value of stock options, as well).

There is a cultural tendency for older people to ignore the ideas of younger ones. They are so inexperienced. And the younger ones want to ignore the sage advice of older folk, because they are so rigid in their ways. Or perhaps just afraid of being supplanted, just as they may have supplanted those who came before them. Communication, working as a real team, can make things better.

You can buy the Faber and Mazlish book from Amazon. Note that Amazon carries books that BarnesandNoble.com claims are out of print (for reasons that can only be censorship). My wife ordered a controversial book from Barnes and Noble, and we first got an email saying that they would have to search for the book. Later, we got an email saying it was out of print. Amazon delivered a new copy of this "out of print" book in less than a week.

Buy Amazon. You will be supporting USENIX sysadmins and fighting censorship by a large company that pretends to be very progressive.

While I am on the subject: don't buy from any Mitsubishi companies until the Mitsubishi Corporation abandons its plans to build a saltworks in the Laguna San Ignacio gray whale nursery (<http://www.nrdc.org>). I think that Mitsubishi makes wonderful large-screen televisions, as well as many other products. But their plans to invade a peaceful cove in Baja California and turn it into an outlet for brine will destroy the last place on the North American West Coast used by gray whales during breeding. Sure, the lagoon makes a handy place to bulldoze drying ponds, but there is no way this activity can be ignored by the whales. I'd like to see some sensibility from Mitsubishi, instead of claims that this activity will be harmless.

On a final note, I finally installed VMware. Several people had written to me about how much they like having it, so I spent the time installing it again. This also involved loading an updated X server, and I am still having problems matching my display and video card, with the side effect of having color-map problems. Still, it is certainly weird to have NT running on a system running Linux. I can scan it, attack it from the network, and I don't have to have two PCs wasting energy just so I can run both Linux and NT. I'll have more to say about VMware once I get the kinks out.

And lots of people have written to me about StarOffice as well. Rather nice that Sun bought it, and I hope they make it better and less feature-ridden than its competitor.