The Murata OKI-78SR DC-DC power converters are a pair of drop-in replacements for the classic linear regulators, LM7803 and LM7805. Unit cost is about four bucks, where the LM78xx is more like $0.60, but they claim 95% efficiency and no need for a heat sink. Sounds like a nice tool to have in one’s parts box.
I have never been a fan of Microsoft, but as the company has become less of an all-conquering juggernaut, I’m starting to see something valuable in its approach to computing which is rapidly being lost in the world outside the increasingly insular Redmond walls.
The last ten years have all been about web platforms, server-side applications, cloud computing: a sort of modern return to the mainframe. Microsoft, on the other hand, continues to be all about the personal computer. Of course they are trying to play in the new world too, but Microsoft still fundamentally sees the network as a service for the PC.
Contrast this with Google’s new CR48 laptop, a device so unstylish, so unremarkable, that it is clearly nothing more than a viewer for the network. The computer is not interesting, Google is trying to tell us: all that matters is the data on the network.
They’re right, of course, and this approach has a lot going for it. Why shouldn’t everyone delegate system administration to the professionals? Why shouldn’t everyone keep all their data offsite all the time, where it’s invulnerable to accidents? Why shouldn’t you be able to get to everything from anywhere, no matter who happens to have paid for the piece of hardware you’re using to get it?
Google, as a company, gives me all the warm fuzzies Microsoft never could, but there’s something wrong with the future they seem to be building, and I can’t completely get on board. The problem with the network is that you don’t own it, you can’t control it, and you are therefore at the mercy of whoever does own it and control it. Once your data lives on someone else’s machine, you have to trust that someone to take care of it, and no matter how reputable, that someone will never care about your data as much as you do.
The Google of today appears to be an extraordinarily intelligent, farsighted, responsible corporation, but organizations change over time as people come and go. Will the Google of ten years from now be as trustworthy? I doubt it. Google is vulnerable to pressure from various governments; this will not change unless they recruit an army and set themselves up as an independent state. Google is not accountable to its users, but its shareholders; this will not change unless they commit a spectacularly unprecedented act of corporate suicide and turn themselves into some sort of user-driven non-profit. Organizations change, but data is permanent, and I am not willing to extend my trust to all the possible future Googles as well as the current one.
I am picking on Google in part because they are the biggest, the smartest, and the most ambitious; they are pushing this new frontier farther than anyone else, and their projects make it easiest to guess where we will all be a few years from now. I’m also picking on Google because they are also the most obviously benign of our grand technological overlords. There are other cloud services, other thin-client projects, and the situation there looks more grim. Amazon booted WikiLeaks off of EC2 the moment they started to look a little inconvenient; they’re certainly not going to stand up for anyone else. Apple, lovely as their design work is, has an even more strongly paternalistic little walled garden going with their iPhone and iPad; I’ve been a Mac user since 1985, but I wouldn’t trust Apple with my data.
I understand the economic forces that are pushing the future in this direction; I just don’t like the direction very much. I have smart friends at these big companies who are working hard to create enormously powerful tools that will change the world for the better, and I wish I could be as wholeheartedly enthusiastic about their work as I believe their efforts deserve. The irony is that the move toward a Web-centric world has cost us the decentralized freedom of communication that the Internet originally offered.
Decentralized peer-to-peer services are the way out of this trap, which is part of the reason governments and organizations dependent on governmental coercion find them so unpleasant. Every central server is a vulnerability, because governments or greedy corporations can interpose themselves as gatekeepers; every service that depends on a central server is thus dependent on their blessing, or at least their indifference.
How can we re-engineer the web so that every node in the network can participate in the cloud system? How can we build something like EC2 or Azure that cannot possibly be shut down by anyone, no matter how large their army or how deep their pocketbook? It’s a hard project, and the economic model is a far bigger challenge than the technical one, but freeing our communication infrastructure from the threat of corporate or governmental interference is the most important project of the 21st century.
Kevin M. and I joined Barry B. for a quick little ski-trip getaway this weekend. We met up after work Friday and drove up to a cabin near Mt. Baker, so we were in line waiting when the lifts started up. The epic drifts of fresh powder we had heard about did not materialize, but it was good snow and the place never got crowded, so we had a great time. We spent most of the day bombing down a single run, over and over, stretching out and encouraging the muscle memory to come back. I missed last season, so I definitely felt rusty, but I had my carving down again after a couple hours, and was working on balance by the time my quads started to give up.
We all got tired around the same time and left an hour and a half before the lifts closed. There’s something to be said for pushing yourself, but there comes a point of exhaustion where your body just doesn’t respond properly anymore, and there’s not much fun to be had after that.
I wore my new ski jacket, of course, and the experiment was a definite success. I stayed warm and comfortable all day, had no trouble moving around, and got a number of enthusiastic comments from the resort staff. The material worked, the style worked, the design mostly worked, and I will definitely wear it again. That said, of course I learned a few things that will help me do a better job next time:
– there is nowhere to hang the lift-ticket tag. needs some kind of loop stitched in along the bottom hem.
– the collar is too snug – it works fine when the jacket is all I’m wearing, but it doesn’t close when there’s a layer of fleece underneath.
– the front flap tends to blow open in cross-breezes: needs more weight or some kind of closure to keep it positioned. Next time I might use a longer zipper and add some side vents, instead of using the front for range-of-motion.
– the hardshell material keeps the wind and water out but doesn’t really keep heat in; the jacket is not much use without a fleece layer underneath. Next time, why not incorporate a microfleece lining?
I’m having a good time with MJ. There’s no real agenda for this weekend – I’m just here to have fun in New York and spend some time catching up with my sister. Yesterday I went up to the fashion district and bought pieces of various fabrics at Mood: nothing too exotic this time, but some interesting, quality material I couldn’t find at home.
After dinner we went rambling around the lower east side and found a place called the Living Room, with a series of bands playing one-hour sets. The group playing when we walked in were a bunch of hilariously earnest music nerds – the front man played a banjo and some circuit-bent kid’s toy – so we didn’t expect much, but the next band was really surprisingly good. Called Feldberg, they are touring here from Iceland, and we’re going to go catch them again tonight in Brooklyn and bring some of MJ’s friends with us.
Today we visited the farmer’s market at Grand Army Plaza and bought some cheese and apple cider. Then we wandered around prospect park and poked our heads in various shops, stopped in at Press 195 for lunch, rode the subway, and wound up spending a couple hours in the Brooklyn Bead Box making stuff. I had just been looking for a couple split rings to repair a cuff link, but MJ got sucked into a Christmas-ornament project, and then I started making a braided wire-and-bead bracelet, and we had a great time hanging out chatting with the staff and working on our creations.
We’re off to dinner soon, then music and dancing and probably a fair amount of whiskey along the way. Should be a good night out!
The code is done and the assembly line is humming. With one exception, everything works the way it is supposed to. I’m not supposed to describe what it is I’ve been building or explain what the project will accomplish, but a descendant of this device will ship as a commercial product someday, and perhaps then I’ll be able to point it out. For now, I’m just glad crunch time is over, and impressed that crunch time at Synapse has been such a reasonable experience. Long days are never exactly fun, but there’s a big difference between pushing hard for a few days, as this week, and having to grind on for weeks or months.
Work. Lots of it.
Shipping soon? I hope so.
Going to NYC to visit M.J. tomorrow evening. Yay.
Hoping to swing by Mood while I’m in the neighborhood.
Cost/benefit analysis of the TSA’s new pornoscanners:
Given the current deployment levels of AIT scanners, these scanners will over a long time horizon only save approximately 2.4 lives a year, while inducing 3.4 incidents of fatal cancer per year over the same period. Even more troubling is the increase in fatalities as people choose to drive when faced with the possible option of an intrusive enhanced pat-down, which can result in a median of 189 additional roadway fatalities a year. All together this means that while the AIT program reduces the incidence of a low likelihood, high risk and high profile terrorist attack, it induces a median of 190.7 deaths across other parts of American society.
I will be flying next week, and I have no intention of going through the pornoscanners. If the TSA is going to force me to submit to an invasive personal examination, I have no reason to make it easier for them, and certainly no reason to accept the health risk associated with X-rays, no matter how minor it is.
Beyond that, I’m tired of the farce air travel has become. What is the point of the fourth amendment if “terrorism” can justify arbitrarily invasive search and surveillance? I am vastly more likely to be harassed and deprived of time or property by agents of the TSA than by any “terrorist” attacker. The fact that the X-ray pornoscanners are, by themselves, more likely to kill you than a terrorist attack is hilarious, but it is true of the entire apparatus: it’s a bogus cure which is worse than the dramatic but extremely rare disease it claims to cure.
In the end, though, we won’t get free of this nonsense anytime soon no matter how ridiculous, invasive, and pointless it all is. The bureaucracy exists, and like any bureaucracy it is primarily in the business of keeping itself in business. The only solution is to avoid the whole mess whenever you can, and gum it up via passive resistance whenever you can’t.
From now on, I’m flying in a kilt.
Ever needed to write platform-dependent behavior in your C/C++ code? Pre-defined C/C++ Compiler Macros is an apparently comprehensive list of platform, compiler, library, and architecture identifiers.
The Freakduino-Chibi is an oddly-named Arduino-compatible board with a built in 2.4GHz radio, plus a software stack implementing the 802.15.4 (Zigbee) protocol. The designer had remote sensing applications in mind, but it looks like a pretty reasonable foundation for the walkie-textie project – lots cheaper than the combination of an Arduino Fio and an Xbee module.
There’s nothing going on with the juggling balls at the moment, but I’ve been banging away at Radian with some success. I’ve rewritten the IO system, which now supports both input and output, as well as providing access to the command-line arguments. It’s now at least theoretically possible to write Radian programs that do something useful. It won’t be long now before you’ll be able to read, write, create, delete, and otherwise manipulate files. Shortly after that I will need to spend some time planning the feature set and development timetable for the first release, since it will no longer be obvious what work needs to be done next. This is a pretty exciting phase of the project!
The ski jacket is still on track to be ready for the beginning of ski season. I’ve stitched and taped both sleeves, though now I’m out of seam tape so I won’t be able to get much further until I can drop by Seattle Fabrics for more. Two and a half yards of the three-layer goretex material turned out to be just barely enough; if I did it again I’d probably get an extra quarter yard just to make layout less of a challenge. Working with goretex is an interesting challenge: you can’t poke pins through it because that damages the waterproof membrane. I laid out the pattern using scotch tape, and have been holding the pieces together for stitching with fine silk pins, put through about an eighth of an inch of fabric in the seam allowance. It doesn’t matter if there are holes in the seam allowance, since I’m going to cut most of it off and seal it with seam tape anyway. At least the fabric doesn’t fray, since the top and bottom layers are bonded to the goretex membrane.
I prepared a talk for Chillstudy yesterday, titled More Than You Ever Wanted To Know About LEDs. The link points to a PDF copy of my slide set. I’m not sure how useful it will be without my accompanying chatter, but at least it should serve as an entertaining demonstration of my facility with Google Image Search. The talk went smoothly, despite all my stressing out about it; the turnout was small, but the people there seemed to be interested in what I had to say.
I’m thinking of putting aside the DJ-hardware project for a while. I’ve had nothing but trouble with the rhythm robot PCBs, and it’s become discouraging. I think I need to let this specific batch of hardware sit until I have had some time to think about what to do next. This would leave me with only three active projects: Radian, the intelligent juggling balls, and the ski jacket.
My ideas about an IO system for Radian are coming into focus. I have decided on an internal architecture; I am still thinking about finding more user-friendly ways of presenting it, but I don’t actually have to solve that problem in order to start building the foundation. I think I’m going to end up implementing something like a
The ski jacket is coming along well – I’ve finished enough of the cotton twill prototype to be confident that the pattern will fit comfortably and offer the range of motion I need. I intend to finish the prototype as a garment of its own, but for the time being I’m going to put it aside and start cutting out the gore-tex.
So I implemented the parallelization system for Radian, and it works.
Yep. Now, uh, on to the next thing, right?
It feels like it ought to feel like a bigger deal than it feels like. Accomplishing a big goal often seems to work that way.
When did the verb “ask” become a noun, as in “their minimum ask was $50k”, and what was wrong with “request”?
Two useful calculators for working with Rc filters:
Electric motorcycles are popping up all over the place. The one that caught my eye today is the Enertia Plus from Portland-based Brammo: 80-mile range, top speed in the low sixties, and the usual torque-like-crazy electric performance. Not only that, it looks terrific and list price – brand new! – is only $9k.
My project this afternoon is a quick little record bag for my DJ costume. Record bag, that is, as in a shoulder-bag that holds actual vinyl records. I already have the big raver pants, and I got this puffy, fur-trimmed hoodie and a baseball hat this afternoon; add a pair of headphones and I’m all set.
But wait, the informed reader may ask: you say you are dressing up as a DJ for Halloween, but aren’t you actually a DJ in real life? And aren’t you, in fact, going to be performing as a DJ on Halloween itself? Wearing your… DJ…. costume?
Ah, yes: but it’s still a costume, because I am dressing up as a completely different kind of DJ. It will be obvious that my DJ outfit is a costume as soon as you hear the music, because I will look like someone who would be playing a completely different kind of music! Obviously.
Also, I will be a zombie.
Maybe it’s the return of grey, rainy weather; my enthusiasm for Radian is back in force and I am making what feels like rapid progress. I built one major component of the automatic parallelization system in one quick burst, and it worked practically the first time I ran it – I guess that thinking about it for so long meant I had a solid design in mind by the time I finally got down to writing the code. Why did I wait so long? Oh well – there’s plenty more to do, and I’m digging through it with gusto.
Tonight I’m doing a little debugging work on the new rhythm robot board. The microcontroller will not respond to my programming attempts – it appears to be dead. I’ve tested all the connections and they seem to be correct. Is it possible that I simply overheated the chip while soldering it in? Another instance of what appears to be exactly the same circuit works just fine in a breadboard. What’s more, I assembled just the microcontroller portion of a second board, and ran into exactly the same problem. It seems strange that I would have fried two chips in a row in exactly the same way, when I had no such problems building the bloom light boards, but the chips worked fine when I tested them in a breadboard beforehand. Oh, well, I’ll work it out eventually. In the meantime it’s tempting to just let the whole thing sit and work on something else, but I really do want to get this gadget running.
This completely awesome home-built searchlight incorporates an arc lamp, powered by a welder, using a trash can as the reflector. I want to take one of these to Burning Man.
a few samples from anti-jokes:
Your momma’s so fat that she should probably be worried about the increased risk of cardiovascular disease.
Why was six afraid of seven?
It wasn’t. Numbers are not sentient and thus incapable of feeling fear.So a Hispanic, African-American, Jewish, and Asian man were walking down the street.
They were involved in a parade that celebrated racial equality.A duck walks into a bar, the bartender says, “What’ll it be?” The duck doesn’t say anything because it’s a duck.
Current projects: intelligent juggling balls, Radian, ski jacket.
I’ve been doing some board layout for the juggling ball prototypes. It’s a tricky problem: we want to stuff a bunch of SparkFun breakout boards into an actual ball and toss it around to see what the LEDs look like and what kind of data we get from the sensors. There’s so little space inside that we can hardly make room for physical wires, besides which it’s just a nuisance trying to cut, strip, and solder all those little bits, so I’m designing a simple PCB that will route signals between the various breakout boards. I have about three and a quarter square inches to work with – definitely a challenge, and so far a most engaging one.
On the Radian project, I’m finally digging into the crux of the system, which is the automatic loop parallelizer. The work I’ve done so far has given me a simple, unremarkable, unfinished programming language with a couple of minor syntax quirks and an unusual internal architecture; now it’s time to make all that preparation start to pay off.
I have moments of doubt, where I wonder whether I really understand this problem as well as I think I do, fearing that I’m going to get hopelessly stuck in the implementation, or – worse – suddenly realize that there is a fatal flaw in the whole plan. When I’m really feeling uncertain I imagine that this fatal flaw will turn out to be something that is old news in the CS literature, but which I’ve never read about nor been clever enough to figure out for myself. Oh, well: even if the venture did turn out to have been doomed from the start, I’ve learned a lot from it, and there would probably turn out to be something else I could do with the codebase.
But honestly? It’s going to work.
1) Sometimes, when I log in, the cursor is invisible. The trackpad still works; I just can’t see what I’m pointing at, and have to guess based on rollover hilights. The only thing that fixes this problem is to suspend/sleep/whatever it is and then wake the machine back up again.
2) Sometimes, when I log in, a message pops up telling me something about a “local network service” that is incompatible with something called an “Avahi” service, and that one or the other of these services – it’s not clear which – has been disabled. I have no idea what any of this means; none of the system configuration programs have anything to say about “Avahi”.
3) The computer does not consistently go to sleep when I close it. Sometimes it does; sometimes it just sits there, screen glowing and fan running, until it either runs out of battery or irritates me so much I go do something about it. This may be an Eee PC hardware problem, but I’m blaming it on Linux because it’s the kind of fit-and-finish detail the Linux people don’t seem to notice.
4) The updater system is complicated, confusing, and doesn’t even work. The “Administration” directory has a program called “Update Manager”, and a separate program called “Synaptic Package Manager”, both of which appear to do more or less the same thing. I might have a better idea what the difference was if either program actually worked; instead, whenever I try to install or update anything, I get a series of error messages about “mirror.uoregon.edu”. What is mirror.uoregon.edu, why does my computer’s OS depend on it, and what am I supposed to do to fix this problem? I have no idea. Thus my package information has not been updated in 356 days, as the update manager helpfully pops up and tells me every month or so, despite the fact that I turned the automatic “check for updates” feature off owing to the above mirror.uoregon.edu problem.
5) Sometimes the network menu doesn’t show up when I log in, and it is impossible to tell whether the machine is connected to a wireless network,or to instruct it to log in to some new network. I have no idea why this happens. Restarting the machine usually fixes the problem.
I actually like this little computer quite a lot, and I think the Ubuntu people have done a pretty good job at polishing up the chaotic wilderness that is Linux. I am just frustrated by design misfeatures which leave me feeling like I am not the one in control of my own computer. Since I have no idea what to do about any of this, I’m posting it in hopes that people who work on the various components of this system may somehow, someday, perform some google search which puts them in touch with this post and thereby gain some insight into issues they might consider working on.
I’ve had a couple of Post Office failed-delivery slips kicking around for several weeks now. The place closes at six, so it’s nearly impossible to get there without making special efforts to rearrange my schedule. Having no idea what this package was, or who it was from, or why I should care about it, the errand just hadn’t seemed that important. Well, this morning I managed to scoot out of the house early, stop by the post office, and pick up these mysterious packages. It was just one package, which they’d tried to deliver it twice; a little tan envelope, addressed from Sofia, in Bulgaria. What on earth? I don’t know anyone in Bulgaria; why is someone sending me a tiny lumpy package which requires signature confirmation?
Turns out it is a pair of FND500 LED displays, which I apparently bought on eBay late one September evening. I really like FND500s; they haven’t been made since the mid-80s, and they light up with a gorgeous deep red semi-translucent glow. They’re not so great for long-distance readability but they look beautiful. I inherited a handful of them from an older electronics hacker’s stash, and bought another couple dozen on eBay once I found out how awesome they were – I periodically check eBay for more but had completely forgotten that I’d actually ordered any.
This looks like astonishingly good fun: Jeb Corliss shows off his favorite wingsuit flights. The bit toward the end where he skims over a meadow and drops into a surprisingly steep canyon at the far side – wow!
Three months ago I posted a list of the projects I had under way. Feeling like I am doing too many things and not doing them justice, I have been working toward spreading myself less thin. I’d like to let the project list shorten as I finish things, then refrain from taking on anything new for a good six months at least.
Finished since last update:
– Handbag for my sister Joanna
– Four more bloom lights, for Eva’s flower sculptures
– Blinking firefly costumes for Chris W. and his group
– Rhythm Robot v2 circuit board
– Flashy silver raver pants
– Blue/grey motorcycle-inspired modified jeans
– Reworked straps on a dress for Nika
Currently active:
– Radian (finishing the initial set of object model features)
– “Orb of Awesome” for Aaron B. (last remaining parts arrived today! finished this weekend, I hope)
– Steadyrocker midi clock device (circuit board layout done, next step is to get it printed)
– Rhythm Robot (updating firmware, designing a housing)
– Custom ski jacket (Martian styling, applied to goretex)
On the back burner:
– Sell the older two of my three motorcycles
– Tune up the Suzuki and install the Corbin seat
– Intelligent juggling balls
– Red shot silk dress w/gold lining
– Ballistic nylon motorcycle pants
– Groovik’s cube dimmer code
– Hammerbox percussion synthesizer
Mars Saxman
2006: January February March April May June July August September October November December
2005: January February March April May June July August September October November December
2004: January February March April May June July August September October November December
Wanderings in Black and Red (previous site)