|
|
perfect work choreography Sunday, January 13 2008
The relatively pleasant weather persisted for another day today and took advantage of the daylight to do necessary outdoor work. Simultaneously, I took advantage of being near-to but away-from my computer to replace its two hard drives with a single one roughly as big as their sum. I would have gladly bought a larger hard drive, but I've developed a policy of spending less than $120 on hard drive purchases, and the state of the art in hard drives limits me to a 500 gigabyte drive at that price. Since at least 1998 I'd been running two hard drives in my main computer, replacing the smaller one every year or two with the largest drive I could get for whatever my price limit was. In the past that limit was as high as $350, but now it's $120. I remember buying a ten gigabyte drive in San Diego in 1999, and in Google I can find accounts of installing a 120 gigabyte drive in 2003, and a 300 gigabyte hard drive in 2006. I should take a moment here to throw some cold water on the "limitless accelerating tech progress" geeks with the observation that commodity hard drive capacity isn't advancing as quickly as it once did. Capacities used to double every year, but if they were still doing that then I could have bought a hard drive of at least 800 gigabytes for $120. (The 300 gigabyte hard drive I bought a year and a half ago cost $115, while the 500 gigabyte hard drive I just bought cost $101. You do the math, because it's a little too non-linear for my brain just now.)
The main reason for consolidating all my storage to a single hard drive was energy conservation, a factor that I didn't consider as recently as 2005. A hard drive uses five to ten watts of power continuously as long as it's on, which can be as much as 24 hours a day. If I can cut that consumption in half by eliminating a hard drive, I'm happy to do it. A side benefit is that computer containing one hard drive is generally quieter and cooler than one containing two. The problem, of course, is that dropping to a single hard drive swims upstream against the current of storage demands. I can get a new hard drive that is the size of my two existing hard drives, but no larger, so the swap gives me no additional space. Fortunately, it doesn't take much work to locate vast gigabytes of crap that can be thrown out, including three or four WINNT folders from long-defunct Windows installations in the various partitions.
I'd had about seven hard drive partitions (which I always name after species of birds), but there is a four-partition limit for a given drive, so I had to consolidate several of my partitions. The actual copying of non-consolidated partitions was easy with Partition Magic; it even successfully made a bootable copy of my Windows XP system partition, saving me a shitload of work. The problems came later during the consolidation of the extra partitions. To do this I used the deeply-flawed drag-and-drop file copy facility of Windows Explorer, a facility whose behavior I have bitched about in the past. Forgive me, but I feel the need to bitch again. The problem I've identified with this copy engine is that it's extremely brittle and fails in the least helpful of ways. If, during a copy, a file can't be copied (its name is weird, it can't be accessed, or it has some deep and mysterious system function), Windows throws up and error and stops. Unless the problem is a question of overwriting an existing file, the copy cannot be made to continue. It dies wherever it was, often leaving a terrible mess for you to sort out. This wouldn't necessarily be much of a problem in a simple copy operation involving a few files and maybe one or two levels of directories. But when you're copying whole volumes' worth of files into a consolidated partition, and when those files are organized into deep directory structures, a failure of a copy can mean that several hours worth of copying were done and several more were left unfinished, and where the copy stopped is on file number N, X levels deep in folder Y. If you want to pick up the pieces and continue, you're forced to drill down to the problem file in the source directory, continue the copy with the next file past the problem one, and then come up a level, continue the copy past this folder containing the problem file, come up a level, continue the copy past... It's fucking idiocy to build a copy system like this! Why can't the operating system continue on its own and wait until the end to present a list of files that, for whatever reason, couldn't be copied? I have never used Windows Vista, but I would love it if someone who has would tell me if this copy crapout problem has been fixed yet. My guess is that it hasn't. I'm sure I wasn't the only one complaining about this back in the days of Windows 98, and of course it wasn't fixed for 2000 or XP. Why would Microsoft have fixed it in Vista? That would be the first sound reason I would have heard yet to "upgrade," though I'd actually prefer a small, smartly-written copy utility, perhaps one that supplants the idiotic Explorer method.
While many dozens of gigabytes of data were being copied by my otherwise-idle computer, I was up on the solar deck sleuthing out my persistent panel leak. During the night the eighty pounds per square inch inside the panel's copper pipe had managed to force a certain amount of fluid through a tiny pinhole leak that I'd apparently yet to fix. This gave me two important bits of information. One: the panel still leaked. Two: it wasn't sufficiently drained of fluid. Unless I could empty it completely, at least along that stretch of pipe, I was never going to get it hot enough to repair the leak. So I slackened the tension cables and pulled out the supports on the panel, lowering it until it was flat. This seemed to have the desired effect, because when I hit the troublesome joint with heat it changed through a series of colors, something it had resisted doing before. But even after that the pressure test failed; it seemed that the pipe had been inserted too shallowly into the fitting and there was insufficient surface for solder to form a complete seal. So I cut a straight junction fitting in half lengthwise two form two curved copper plates and used one of these as a sister beside the problem fitting. This provided additional surface area for the solder in the joint to grasp and flow over. Happily, this seemed to fix the problem.
During the hours of pressure testing, I removed the last of the plastic greenhouse sheeting that had been the panel's cheap-and-easy cover plate. Then I pulled out all the nails that would be in the way of the glass and sanded away the glue to make a nice flat place for the glass to lie. When I was sure that the leak problem was solved, I cleaned the collector plate and pipe and repainted the parts where the black paint had burned away during my soldering.
Periodically I'd take breaks from the solar panel work to collect firewood in anticipation of a snowstorm that was supposedly on the way. Lugging large pieces of oak up to the woodshed was brutal work, and to a lesser extent so was the task of dragging in a number of small dead White Pines to be used as kindling. When Gretchen cooked dinner tonight, I ate like a farmer. All that climbing and carrying had left my body feeling as though it had been beaten up.
For linking purposes this article's URL is: http://asecular.com/blog.php?080113 feedback previous | next |