I’ve been a fan of the Evil Mad Scientists since I first saw Egg Bot back at the first Maker Faire. Windell and Lenore essentially invented the type of business they run now (along with Adafruit, Sparkfun and others) and the two of them are finally doing Evil Mad Science full time.
I love it when smart people can grow a totally different business into something that can both support the community and the people behind it. Congrats to them!
I don’t often recommend reading company handbooks but the Valve company handbook is definitely an exception.
It feels like a bunch of smart nerds got together and ignored all the management advice they were given from every other boilerplate company. Then they made a bunch of awesome games and didn’t compromise on their values. Now they’re rich smart nerds.
via @rr
Possibly my favorite part of this montage of dumb things filmed with a high speed camera is the physics.
At 2500fps it gets really obvious that gravity and air resistance play a major role in slowing down fast moving things. Even small pieces of watermelon that are ejected by fireworks get noticeably slower right after the explosion. I won’t bother with the kinematics, I’m sure you get all the video analysis you need from Dot Physics.
I noticed two really fast things in this video, one of which is obvious and the other not so much. Can you guess what they are?
Here’s a pretty awesome photo from last night’s rare thunderstorm in San Francisco. It shows lightning simultaneously striking all four western span towers of the Bay Bridge.
Mother Nature: more like this, plz!
Update: Another shot of the same thing!
There’s something about this tiny, open source printer that makes me think about non-digital communication in a positive way, which is a somewhat rare occurrence. Something about the way it’s designed and how nicely the print quality is… and the simplicity of it all.
Also, when I said “open source” I really meant it. You build the printer.
via @atduskgreg
I haven’t read Design Is a Job yet but you should buy it if you have clients, even if you don’t specifically do design.
When he can tear himself away from Twitter, Mike Monteiro does some really great work. In terms of work for his clients and when talking about his process on the Mule Design Blog. And that’s why I can recommend Design Is a Job without reading it.
There are a lot of design decisions that I’m not too fond of in Ice Cream Sandwich but the Android design pattern document is pretty damn good at not hiding anything behind secret gestures, unlike Metro and that new iPhoto app.
I think I’ve linked to the whole thing before but they just updated the settings patterns. They’re clear and consistent, though I don’t really like the List > List > Modal pattern they use for options that require more than a checkmark.
Here’s a fascinating look at some unique problems our society is experiencing during the transition between primarily Human to primarily Computer-based control of critical systems.
“People Make Poor Monitors for Computers” describes the interesting case of Air France 447 which experienced some loss of instrumentation data while flying through storms near the equator. Control of the aircraft was ceded to the Human operators in the midst of an airspeed indicator failure and the operators couldn’t make sense of what was going on (and in fact made very wrong choices) before it was too late to recover.
The article doesn’t provide much in the way of solutions but it does break down the problems inherent in control systems that are only controlled by Humans when something goes wrong. The TL;DR version is essentially “complex computer systems have lots of assumptions built in” and “humans are really terrible at switching from monitors to controllers of complex systems”.
Our intuition tells us that computers aren’t good for problem solving so we design these systems with human fallbacks. And it made perfect sense to design this way as we entered the era of computer control systems. Computers were hardly proficient enough to control everything so we handed off a decision here and there.
Eventually (i.e. now) control systems are so complex that failing over to Humans has become a bad idea. Unfortunately we’re still of the mind that computers can’t be fully trusted with our control systems.
Notoriously, the Space Shuttle was capable of making all of the decisions during launch and landing except for lowering the landing gear. Furthermore, the approach for landing could be mostly automated by a system with a reaction time at least a few orders of magnitude better than a Human but has always been Human-controlled. Even enabling Computer control of the landing required an extra cable strung through the main cabin to connect two physically distinct control systems, supposedly purposefully designed to discourage its use.
And maybe it’s a valid point that Computer control shouldn’t be currently trusted. After all, we design the control systems to be handed off in the event of an emergency rather than try to make sense of a rare and possibly unpredictable event. Hell, we design the control systems in the first place.
Or perhaps it’s simply an issue of design considerations. The only way we’ll develop better control systems is by depending on them in unpredictable situations.
Google takes an interesting approach to control with their self-driving car. If the control system detects a failure that prevents it from continuing on automatic pilot, it warns the driver that manual control will resume in a certain amount of seconds so the driver has time to evaluate the situation before getting dumped into control.
Your basic car has a fairly limited set of systems that one has to be aware of to decently drive, so this doesn’t seem too dangerous. However, luxury cars are trending towards more computer monitoring systems such as lane-awareness, distance sensors, turn-aware headlights, etc.
It’s not a stretch to expect further automation, particularly as cars that drive themselves become available for general use. Pilots are specially trained yearly (even when they’ve been pilots for many years) to deal with a range of unexpected situations aboard aircraft. What happens in 20 years when a teenager who has never had manual control of a car (except when practicing for their license test) is subject to a LIDAR failure and has to take control while on the highway?
It seems unlikely that the regulatory bodies in place now would allow this sort of thing to happen without an incredible improvement in systems control technology. And we all want self-driving cars, right?
Yes, this is all speculation on the future but I hardly think it’s beyond the bounds of likelihood in the next 10 years. It’s more proof that we need to start developing and depending on computer systems that behave well in unexpected situations lest we subject ourselves to the limitations of our own Humanity.
I’m not entirely sure what this giant air gun is used for (“research”, clearly) but the people who built it made an excellent video showing off its capabilities with slow motion.
I’m sure there’s a good reason to launch things at high speed with “low” acceleration (< 600g) but the author of the video is clearly playing it casual. From the description:
In the world of ordnance system development the need for system and subsystem testing under dynamic conditions is obvious.
Obviously.
The creator of Minecraft has released the basics around his new game, 0x10c. It’s scifi-ish, based around space ships and all the great things that come with it.
If Notch taught a bunch of kids about circuits and logic gates in Minecraft, he aims to teach them about programming in assembly as each ship contains a simulated 16-bit CPU. Here’s a spec for the simulated CPU.
It’s time for another episode (yet again) of A Technical Guide to a Niche Topic! This episode is sponsored by my insistence that personal research not be lost to the void.
Last weekend I spent an inordinate amount of time getting Ubuntu running without a monitor on a mac mini so I figure I’d write up my findings. FYI, this is a 2009 Core Duo Mac Mini. It’s not the newer unibody design though I expect many of the details still apply.
The hardest part of the whole damn thing is getting the mini to boot into Ubuntu at all since all new Macs boot with EFI and support for anything other than BIOS is rare for Linux. Luckily Boot Camp does the hard work for us since it’s designed to run anything on that partition in a compatibility BIOS mode.
A fresh install of OS X would be great, but an existing install is fine as long as your disk isn’t too full. To maintain our Boot Camp compatibility mode we’re going to keep a small OS X partition around forever. We’ll can get potential Apple firmware updates this way too.
We start in our fresh (or not) OS X machine. Download and install rEFIt, a tool that lets you choose your boot disk when you start up. You will not see the rEFIt menu the first time you reboot. It takes two reboots, don’t ask me why - but don’t reboot yet.
Open up the Boot Camp utility and set the OS X partition to something small (~10gb) and set the rest as the Boot Camp partition.
Insert your Ubuntu installation cd (yes, physical media) and reboot. Twice. The second time your rEFIt menu should show up and a smug penguin should be holding a CD. Choose that option.
Go through the Ubuntu installation as normal until you get to the disk setup. Choose to manually partition, then format your big empty Boot Camp partition as ext3 and don’t create any more partitions. Ubuntu will probably hate on this setup as you’re writing it to disk, but it will let you do it. Go through the rest of the ubuntu install.
You may be able to add a swap partition later but it’s not required to boot and run Ubuntu. I’m running mysql, php-fpm, nginx, etc. and I’m only at 20% RAM usage (of 2gb).
The machine will restart after installation and rEFIt should show smug penguin again but as a disk this time. Boot it! Everything should come up OK… assuming you have a monitor plugged in.
This is probably a good time to change the rEFIt config to tell it to boot Linux first, particularly later when you won’t have a monitor. You’ve got to be in OS X to do it, just edit the config in /etc/EFI/refit/.
I wanted this mini to live somewhere hidden away without a monitor so having a monitor constantly plugged in was not something I was interested in. And, of course, without a monitor plugged in the system was never getting to Ubuntu.
After restarting a number of times with screens plugged in or not, I determined that grub was halting the boot for whatever reason. To the googs!
Failure again. I set GRUB_CMDLINE_LINUX_DEFAULT="nomodeset" and a few other grub options, updated grub and rebooted a few times before being unable to fix the issue. I had to go to plan B.
When searching for headless Ubuntu boot issues, I came across an interesting solution that people running folding@home devised. According to them, most systems don’t power a second GPU (or a first for that matter) if a monitor isn’t plugged in. They wanted to run the folding stuff on an alternate GPU so they had to come up with a method for powering that GPU without actually having a second monitor.
The solution is stupidly simple. VGA connections can “simulate” a monitor by having resistance across 3 sets of pins. This can be done by literally inserting resistors into a VGA connector, as shown below. It’s the DVI-to-VGA connector that shipped with the mini.
I searched around for exact specifications and most people say 75 or 68ohm resistors (purple-green-black or blue-gray-black). I only had a few 68ohm resistors on hand so I used those. Apparently anything from 50-150ohm will work (not tested, explosions lol!).
Now grub loves me and restarts jump right into Ubuntu. I tucked the mac mini away and it has been happily chugging away since. Thanks, internet!
Ah, more dystopian future movies. Not quite a remake of 1990's Total Recall and possibly more closely based on Dick's We Can Remember It for You Wholesale.
There are just too many awesome ways to go with the memory implantation theme, this particular story deals with being able to completely wipe and implant memories, of course. Believe it or not, I'm not sick of the what-is-real-and-what-is-not theme that's been ridden so hard for the last ten years. Is there some unspoken rule where movies dealing with reality must be good? Keep 'em coming.
The second theme that is noticeable (and enjoyable) is how Quaid (played by Colin Farrell) is subconsciously alerted to his condition by strange and violent dreams. Mother nature is able to subvert even the most complex human technology. I wonder if this theme will play out in the end of the film or if they'll stick with the alien artifact from the first film.
source
