I’ve been, unsurprisingly, taking a lot of pictures of the baby recently. My dad visited a few weeks before the baby was born, and showed me how to adjust the white balance on my camera, and a little of what changing that can do. Pictures of the baby have made it even more clear to me how important correct white balance is to good pictures.

Here’s a striking example:

Here is a photo I took with the default white balance setting:	Here’s that same photo, corrected using Aperture’s white balance “shade” preset:	And here’s the photo I took immediately afterward, with the D5000’s “shade” white balance preset:

What a difference! Adjusting the white balance in post-processing is cool, but good techniques (and good tools) still really matter.

My employer uses Cisco’s VPN client for most remote network access. It’s a fine tool for many purposes, but I’ve found it increasingly frustrating over the last year. It has to be manually started, it kills TCP sessions and complains to the GUI when I put a machine to sleep even for a moment, and worst of all its kernel extension causes more crashes on my machine than anything else. Time to replace it. Fortunately, we also have an SSH tunnel system. It’s not good for anything but SSH port forwarding, but port forwarding is all we need.

We want an SSH tunnel that will run whenever there’s a network, turning on as soon as I log in and restoring itself when the machine wakes up from sleep. Apple’s launchd is designed for exactly these needs.

more

Robert Graham writes that cyberwar is fiction:

“Cyberwar” and “cyberweapons” are fiction. The conflicts between nation states in cyberspace are nothing like warfare, and the tools hackers use are nothing like weapons. Putting “cyber” in front a something is just way for people to grasp technical concepts, the analogies quickly break down, and are useless when taken too far (such as a “cyber disarmament treaty”). Unfortunately, it’s the clueless people who believe in these analogies that are driving national policy.

So far, so ordinary anti-cyber screed. Graham is right that analogies have limits; that all abstractions leak; and that “cyberweapon” is a word that describes very little. But then he moves into believing a popular myth about the military—one so popular that many in the military believe it themselves:

Moreover, the military is very goal driven. They want weapons that have a specific effect. That’s not how hacking works. Hacking is opportunistic. For example, let’s say that you want to attack Iran. You might give your cyberwarriors the task of taking out their radar. That’s not something the cyberwarriors could do: chances are good that the exploits they have will have no effect on Iranian radar computers.

This, and the associated bigger-is-better idea, certainly exist in many government acquisitions shops. That’s how you get the F-14, a plane so big it has to be folded to be put away. That’s the approach chosen by every leader whose effectiveness can be measured as a scalar. A strategic bombing campaign is measured by its destruction; strategic bomber commanders therefore want big bombs, many bombs per bomber, and many bombers. More bigger weapons will give more effectiveness at a narrow mission.

But there are already many more complicated missions undertaken by the military, missions whose success cannot be judged on a scalar. Recent events suggest counterinsurgency, a complicated mission with no clear measure of partial or graduated success. We don’t even have to look that far: fighter planes are not measured on such a scalar. A fighter plane is effective to the extent that it can fight well against enemy fighter planes. There’s already a theory of how to compare fighter planes for effectiveness. Our plan for network defense and offense needs a similar theory of maneuverability, of excellence in competition with others, and of weapons that exist for no purpose other than to aggressively oppose the weapons of our enemies.

If the DoD is ready for a Cyber Command—and it wasn’t ready for an Air Force for decades after the introduction of military airplanes—this is what it ought to do. Not vulnerability management, though someone has to do so. Not exploit development, though that’s a good project for the industrial base. Any Cyber Command should hold as its first priority the development of an effective discipline for weapons in a networked world, the comparison of one weapon to another, and the assembly of a force able to stop other groups from doing what they want—not damaging stock markets or disabling power stations, but frustrating the national and quasinational (e.g., Nashi, Basij) network aggressors.

One of these things is not like the others. One of these things is not the same:

• Ask Alice to…
• Be drawn into…
• Call for…
• Disavow…
• Lift…
• Negotiate…
• Oversee the flow…
• Pledge…
• Press Bob to…
• Recognize…
• Represent…
• Revoke its charter.

Which one of these things is not like the others? Which one of these things is not the same?

ready for my answer?

Names are important. Your name (or at least, my name) tells a story about who you are, so it is no surprise that Major Life-changing Events often change our names.

As an example, I will trace the history of my own name, since it’s the one I know best:

When I was born, I was named Katherine Hanley Allen, but called Katie. Katie was my mother’s vision for me: she knew she was going to have a daughter, and that was her daughter’s name. She wanted me to have a long name, though, so she was considering Kathleen, to honor her mother-in-law’s Irish heritage. Several other women in my family were Catherine or Kathryn or Cathryn, though, so she compromised to Katherine, and I was given my grandmother’s maiden name, Hanley, as my middle name.

When I went to college, my name changed—not legally, but through a sequence of accidents I went from being “Katie” to “Kat”. That changed how people interacted with me. My family mostly still calls me “Katie”, or sometimes “Kate” (or, in one inexplicable case, “Katrina”) but my friends, colleagues, and in-laws call me “Kat”. (Some of my colleagues call me “Katherine”, which is in itself a reflection of how we interact—a bit more formally than a very very short name would allow!) Occasionally, a stranger will call me “Kathy”, which automatically loses them some of my respect—it’s not my name, so they lose points.

My name changed again when I married: I appended my husband’s surname to my name, turning mine into a middle name, and became Katherine Hanley Allen Sniffen. It serves to reflect the change in my life: still myself, but now part of a new family.

And now, with the birth of my daughter, there are new names all around. She is named for my great-grandmother Mary Emalene, the only girl of 9, and for Brian’s great-great-aunt Amelia, and for the goddess of Memory, Mnemosyne. But Brian and I get new names, too—we are still ourselves, but we are also “Mommy” and “Daddy”. That may be the most dramatic change of all.

Our daughter was born on Sunday, May 30, at 9:32 AM

Emilene Mnemosyne Sniffen
8 lbs, 15.2 oz
20.5 inches long, with a 37 cm head

We spent 48 hours in the hospital recovering, and now are at home with the most adorable baby in the world!

I had the luxury of learning about computers at a very particular time in history. The Apple IIe was released just in time for me to learn to type on one. The Commodore PET, with its built in screen, modem-tape storage, and tempting flyback, was obsolete just in time for me to take a few broken machines apart and build a new one.

Linux as an amateur workstation environment matured while I was in high school and college, with plenty of free time to re-install, to experiment with new distributions, and to learn by breaking things. The Linux 2.0.18 kernel was plausibly readable by an interested amateur. It’s less clear to me that all the facets of the 2.6 kernel so readily reveal themselves to a contemporary amateur. The construction of a world happens only once. Some future generation will see the construction of each technology I cannot imagine—but how can we teach those who are here now about the tools we have now?

I’ve been facing a related problem with friends who don’t have quite the same technical background I do: they lived through the same time, but they weren’t in the right place. Those of us who were fortunate enough to be looking in the right direction learned magic Jedi powers: to decode HTTP, TCP, IP, and maybe ASCII by eye; to walk the many dimensions of the OSI model; to live a life free of spam and indeed of online advertising; to run on computers only those programs we wish to run, and no others; to replace the careers of mortal men with very small shell scripts. How can these great acts be performed by those who have not learned how? And once a child has grown, how should an adult learn these things?

My advice comes from the most famous teacher of one grown too old to learn: Yoda! Yoda teaches Luke well after Luke was too old to learn Jedi tricks by conventional means. Yoda teaches skills that are meant for interacting with others in a high-tech future-fantasy environment. But he teaches them alone, out in a swamp without any technology at all. He teaches Luke how to be free of slavery to his emotions, how to tell real from unreal, witchcraft from real knowledge, and how to be a Jedi. Those who wish to learn to really use computers may benefit from joining Yoda in his swamp.

Would you like to be a Jedi?

Throw out your Windows machine. Lock your Mac in the closet. Get rid of everything with a friendly and helpful GUI. The GUI makes it easy to manipulate anything you see—but it makes it impossible to manipulate anything you can’t see. And the GUI, because it loves you, shows you only lies.

To be free of lies, you must be free of the friendly GUI. So lock it up outside the reach of your temptation. Now you’ll need to make a computer. You don’t have to make it from scratch—you only have to assemble the last step. Take a piece of hardware that you can treat as blank, and install a basic, non-user-friendly Linux distribution. Debian, Gentoo, and Linux-From-Scratch are acceptable choices. I don’t believe Red Hat ships anything in this class any more. Ubuntu is right out. Let’s assume you pick Debian, because I would.

Now use it to live your life and do your work. Read your e-mail. Browse the web. You’ll find reason to write some programs after a while. There is a GUI, but in general you’ll be happier if you avoid it and especially any part labeled GNOME.

The system includes extensive documentation at http://www.debian.org/doc/. Every software component installed is documented in /usr/share/doc. You’ll be so well supplied with this good documentation that you’ll never miss the banned documentation.

Oh, I didn’t mention that, did I? Nonetheless, the most important rule of this collection of advice: use no technical method or instruction found on a web forum, wiki, or mailing list. There’s a place for those later, when you have taste and wisdom. But at the stage of education you’re at now, those are dangerous. They’re witchcraft: ignorant people with half-understood cargo cult ideas explaining to each other why John Frum, American, will bring new cargo if only they build an ersatz airfield. They’ll tell you that to fix some problem, you should change some configuration file—but they won’t have read the program that uses that file, nor understood the hardware it controls. And until you have read that program and understood that hardware, you want to live with the problem.

So use all the included documentation you like, use all the official published documentation you like, and ask anyone you like for explanations of how things work. Ask me, if you like. When you’re ready to use witchcraft advice, you’ll know—and you’ll be done with this exercise, ready to leave Dagobah for whatever reasons seem right to you.