Saturday, November 01, 2003

Telirati Newsletter #11

This is one of the best essays I have ever written. In it, I reject the idea that the complexity of PCs can be managed away. Just read it. True as ever.

Newsletter #11: Live ammo

The armed forces go to great expense to train people in the use of deadly weapons. Misused, mismanaged, misapplied, or poorly understood, weapons kill more of your own than of your adversary. After all, your troops are at closer range than your target. The army takes recruits who are generally teenagers, and, through enough training, make it possible for them to carry infantry weapons and operate heavy weapons without causing carnage on the training field. Military training stems from the nature of the job and, pivotally, from the nature of the tools of the job.

In my two previous newsletters, I touched on the subject of PCs having overrun their natural constituency. PCs are more successful than they should be in an ideal world. PCs are general-purpose computers built in the image of their minicomputer ancestors. They give the user unlimited flexibility and unlimited ability to make mistakes, far outside the scope of any other kind of system people use. The PC, when deployed as a standard-issue office accouterment, the way a rifle is issued to every infantryman, is like passing out an automatic weapon with live ammo. PC users get nothing like the level of training that the teenagers with the real live ammo in their clips get.

The result is carnage: People make mistakes. They blame their tools -- the PCs. They blame the people charged with keeping their tools useful – the MIS department. They come to fear or loathe this very useful tool. Or they worship it and supplicate before its priests like some cargo cult. To add to the generally unsatisfactory situation, the relationship between PCs, the users of PCs, and MIS departments is very unnatural.

Before there were PCs, there were mainframes and minicomputers with terminal attached to them. The computers were operated and maintained by a cadre of trained technicians. In those historic times, computers and telephone systems were remarkably alike in the way they were installed and operated. Dumb terminals (“very thin clients,” if you are looking for capital to build them today) connected to the computer and phones connected to the PBX. This diverged when some users of the computer – the “power users” that mattered most in justifying the cost of the computer – became dissatisfied with the lack of responsiveness from the MIS technicians to their needs. Telephone users, by contrast, never had such a beef.

The power users were, generally, influential in the organization, effective, and intelligent. They got what they wanted: computers on their desks entirely under their control. They thrived. They analyzed hundreds of financial outcomes while the PC-less drones waited for a single report. They created documents online. They stored their own databases. They became even more powerful, effective, and insufferable. Everybody wanted to be like them. Everybody wanted (or at least made as if they wanted) a PC.

And the MIS department still wanted their jobs. So the old guard made a appalling deal with the masses: They would all get PCs, just like the revolutionaries who stirred up all this trouble to begin with, but MIS would undertake responsibility for making PCs safe for non-revolutionary people.

It is as if the army, having seen the success of rifles in a pilot program in the Special Forces said “Let’s give everybody a rifle and ammo” and omitted to train the soldiers except to tell them to “point and click.” Instead they would use “rifle management systems” to try to maintain central control over rifles, and might commission arms makers to create “zero administration rifles,” or perhaps “network rifles” in which only officers control the triggers. It is pretty clear this won’t work with rifles, and it works not much better with PCs.

How does a teenage soldier manage his rifle? How is it that his buddies aren’t constantly being treated for accidental wounds? Why is there no need for a “rifle help desk?” The answer is that the soldier is intimately familiar with his rifle. The soldier knows his rifle, literally, inside and out. He could take a pile of rifle parts, rusted, pitted, wet, and clean, repair, and assemble them into a perfectly working rifle and do it in the dark with a battle raging around his ears. The rifle, like the PC, is not an appliance. It is a general-purpose tool, and a very powerful one.

In a business culture that is overall very fond of military analogies, in which concepts of unit cohesion, tactics, and strategy are borrowed from modern and ancient military thinkers, the PC, the business infantryman’s rifle, has thus far gotten short shift. This is due to the fact that senior management is just now getting past the stage where a computer is either forthrightly banished from the CEO’s office, or sits on his credenza as a decorative item. A senior military officer that is so out of touch with the “grunts” is a clear sign of trouble, but in business, this is the rule, not the exception.

There is also the fact that, for all the talk of “empowerment” in business, the typical PC user is fairly powerless and confined. Open book management, flattened organizations, and internal entrepreneurship are going to have to take deeper root before many PC users can make the most of the power of the systems at their fingertips. But even those that have the level of access and autonomy to make full use of a PC are hampered by training that, if present at all, is limited to pointing and clicking. What if the typical PC user had the same level of familiarity  the almost matrimonial familiarity  that the soldier has with his rifle? You would have an employee as deadly to his adversaries as a well trained soldier is to his.

You, in your organization, have the opportunity to make your colleagues and employees as familiar with their PCs as infantrymen are with their rifles. If you are typical, you are awash in old PCs  486 PCs with too little memory and disk to be useful. Take a handful of these old clunkers. If you have to buy them, they should cost about $300 apiece. Dismantle them. Buy some copies of a book a PCs, of which there are several, that shows how to build up a PC from parts. Give a handful of your people copies of the book, an up-to-date Windows installation disk, and a dismantled PC. The course should be largely self-directed and self-paced, but should be guided by a person who has already built up a PC from components and can steer the participants to the sources of information they will need to answer questions.

The result should be a group of people for whom the PC holds little mystery. And there is every indication this should be broadly applicable. Building a PC is easier than overhauling a car engine, plowing and planting a field, building a shed, or fixing a leaky pipe. It is easier than learning to play a pennywhistle. It requires no talent or dexterity or practiced moves. The downside risk, which is a “smoked” 486 motherboard, is easily absorbed. What is the advantage of knowing one’s PC from the inside out? The advantage is that, even though Microsoft has done a brilliant job of hiding the complexity and the quirks of the PC architecture form users, the basic nature of the PC still comes through, especially when something goes wrong. With a couple hundred dollars more in hardware, these training PCs could be networked by the participants in the training program, and even more mystery would fall away from the PC.

This all started with the thought that the PC is not for everyone, and it has become something everyone has, and that this imbalance is the root of a lot of evil, especially in the way larger organizations manage their PCs. One way to solve the problem is to look to simpler computers. The larger Windows CE systems will address this need. But we will be cranking out tens of millions of new PCs every year as far as they eye can see. And they will have almost all the legacy baggage of today’s PC (some will begin sloughing it off). The best solution for the world as it stands is to try something radically new to raise the level of confidence of the average PC user. A lot of people arrive at this level of competency and confidence on their own. The sales of books on how PCs work attests to this. As long as PCs remain, inherently, the tool of the power user, it looks like a better bet to upgrade the power of the user than to try, through management mechanisms that were never designed for the PC in the first place, to make the PC “safe.”

Copyright 1997 Zigurd Mednieks. May be reproduced and redistributed with attribution and this notice intact.