What Would You Tell Leonardo?

In a 1992 book of cartoons by Sidney Harris titled "Chalk Up Another One," one of my favorites shows two people walking down a city street. Theyre surrounded by modern artifacts—cars driving by, a jet overhead, a computer in the store window that theyre passing. One person says to the other, "Look, I would say to Leonardo. See how far our technology has taken us. Leonardo would answer, You must explain to me how everything works. At that point, my fantasy ends."

There are people who maintain that if you have to know how something works, the technology is fundamentally flawed. I agree with that sentiment only up to a point. Almost all the time, Im happy to see a piece of technology that works so well that Im free to ignore its inner details. That kind of user experience implies both good design and superb implementation.

But call me greedy. In addition to having things work without my constant low-level intervention, I also want to have the low-level innards visible at a minimum, controllable when possible, and modifiable by anyone whos willing and able to do that work.

I demand some degree of transparency because all artifacts have failure modes. If you cant see into the box, you cant even begin to diagnose the problem. I think Ive used this example before, but a Google search cant find it so perhaps most readers have either forgotten it or not previously heard it from me: If you cant start your car, youll do one thing if the engine is at least turning over (probably a fuel problem) and do something else if the engine wont budge (probably a dead battery).

My "whats inside?" mind-set leads me to keep some diagnostic monitor on the screen of almost every PC that I use. On the Sony Vaio laptop that Im using to write this column, a row of Norton System Doctor icons at the bottom right of my screen is warning me that my physical memory is pretty low. Hey, its a 5-year-old machine: 256MB was a reasonably generous configuration back then, but now its helpful for me to see when Im likely to start thrashing the disk with virtual-memory paging.

A users minimal understanding of memory paging can make the difference between "my machine is too slow" and "my machine needs a memory upgrade" or even just "I need to keep fewer browser windows open at once."

On other machines, I rely on other indicators. My Apple PowerBook G4 has a lot more memory, so I usually keep only a CPU workload bar-graph display at the lower right of my screen. Ive learned some interesting things as a result. For example, in Mozilla 1.7.13, merely holding down the mouse button on a windows vertical scroll bar "elevator box" will max the CPU load, while Firefox 2.0 and Safari 2.0.4 dont care about the button state unless Im actually dragging the box to scroll the page.

Given the amount of time that I spend using browsers to peruse lengthy HTML documents, this certainly affects my choice of weapons as I try to keep my lap cool.

The cue doesnt need to be visual. On the Dell XPS in our family room, a gigabyte of memory means that theres almost always a slice of several hundred megabytes free for the systems scratch pads—and thats how I like it, thank you. But the pitch of the cooling fan, believe it or not, can tell me as I walk by whether one of my sons is really doing homework or actually has a game running in a minimized window—still burning CPU cycles while it waits for Dad to move on.

Speaking of moving on, I said last week that Jan. 22 would mark the end of my time with eWeek. Thank you for letting me be part of your world view for these past 18 years—and make sure that your view continues to include clear cues for when somethings about to break, with a strong set of options for what to do when that warning comes. Someone out there is counting on you to know how everything works.

Former Technology Editor Peter Coffee can be reached, for a while, at posteweekpc@earthlink.net.

Check out eWEEK.coms for the latest news in desktop and notebook computing.