Thursday, April 10

I rode in to work today for the first time in months, along the foothills down 36 with trucks full of debris and construction materials rattling past, out onto the plains along Nelson Road. Everything has that sort of springtime muzziness about it, the ragged-edgedness of early grass through late-winter grime, the smell of thawed-out earth, a haziness in the air, a certain texture to the wind.

The edges of the state highways and county roads are littered with wintertime gravel and gouged in places from heavy equipment. In fact the scars and disruptions from September’s floodwater can be read all over the landscape, if you’ve got the vocabulary. I think a lot of the people on these roads as the tourist/cyclist/daytripper season starts don’t have the vocabulary.

So it’s late on a Thursday night in April. I’ve been rolling in and out of a heavy sleep for hours now, thinking/dreaming about internet security, Science Fantasy, the sensation of typing, family, code, weather, religion, desperation, dirt.

I’m physically exhausted, in that way that only accumulates across weeks or months. I’ve been working too much and accomplishing too little, going out too many nights and not making it back to the town where I live for days on end, operating on an excess of coffee and irregular restaurant food and torrents of beer. I’m badly dressed because the laundry got ahead of me. As usual, the warm-season transients moving back into Boulder sometimes take me for one of their own.

The other day this guy, a put-together enough looking guy with a good backpack and clean clothes and an expression I mistook at first for sanity, asked me if I was Gabriel. I was rounding the corner into the bus station down on Walnut, and I’d nodded at him sitting on the bench with his pack, and when he asked me this I thought he meant was I somebody he’d met around here. When I said no he told me he’d seen Gabriel in Wahoo, coming out of the wall of the sky in a Polaris-class space ship, or something to that effect, and I realized he meant something a lot more quasi-Biblical in nature. I just kept going towards my bus. As much outright madness as I have encountered in my life, I maybe have less idea what to say when it crops up than I ever have.

The internet broke earlier this week. An important, widely used encryption library turned out to have a bug which leaks memory on the target machine to an attacker.

Every time this kind of thing happens, you get a set of people saying, in essence, “hey, the math works, this is just a software problem”, which has long struck me as being both true and worse-than-useless. I have a real hard time not reading it as “hey, the math works, so it doesn’t matter to me that people can’t trust or even understand the use of encryption”.

It’s not that I don’t grok the distinction between abstract concepts and concrete implementations. If you can’t draw those lines real clear, you aren’t fit to write much code at all, even the kind of glorified clerical apparatus I’m responsible for in my working life. It’s more that, well, encryption is completely fucked for actual human use, and this isn’t some kind of incidental. It’s the whole problem. We haven’t yet figured out how to build secure systems, most of our infrastructure is compromised by various malefactors, privacy is a dead letter, and surveillance is now close to ubiquitous anywhere this side of the industrial revolution. Actually-existing-encryption as a remedy for this situation feels almost hopeless.

I’m coming around to thinking that this has a lot to do with the nature of the underlying thing. The implementations aren’t just broken and confusing because programmers are fallible and software is hard (though those things are true). They’re also broken and confusing because, as much as perfect knowledge of the world is impossible and everything is shot through with secrecy, information will out. Technologies deliberately situated in the domain of entropy and set against the grain of all those leaking bits - well, maybe you’re just going to have a bad time.

p1k3 / 2014 / 4 / 10
tags: topics/colorado