Wednesday, August 28, 2013

Raspberry Pi Micro MiniMac almost solves my problem (why digital data isn't always good for archiving)

This could ALMOST solve a recent problem we had here at Digital Diner.
A fellow named John has built a magical little masterpiece.  It is a Raspberry Pi-based micro Macintosh.  It is a 1/3 scale replica of the original Mac... you know, the 1984 one.  It actually runs Mac OS 6, which is quite amazing.  It is quite small and quite cute.

It is a great piece of work and the result is pretty impressive.  The case is packed with a Raspberry Pi, a power supply, a small display and a USB hub.  While the disk drive is not real, this Mac has a color display, USB and Ethernet, which the original didn't.  Plus it is based on a $35 computer... I think my original Mac was $1800 (in 1984 dollars).

Now to the problem that this almost solved.  A while ago, one of Monika's advisors contacted her to see if they could get some of the data from her PhD thesis.  Now I don't think that data is, um, well, really old.  We tried to figure out where it might be and decided that it was probably on an old Mac SE in the garage (see? you should never throw anything away).  Well, I pulled it out and fired it up and after many years of being ignored, it fired right up.  There was a Y2K bug, so I had to set the clock back to 1999, but otherwise the system just worked.  Strangely, it seemed to boot much faster than any of my modern machines.  After taking a little trip down memory lane and then wandering through the disk drive for a bit, we were able to track down the data that had been requested.  So far, so good.

This is where the problem began.  You see, we found the data file and could even fire up a program to look at the values, but the next step was a bit of a problem.  How do we get the data off this old Mac?  I found some old floppies, so I could copy the data there, but then what?  I don't have any way to get data from a 3.5" floppy onto a modern machine.  Maybe I could connect it to the network?  Nope.  The only network this thing supported was AppleTalk, which was a glorified serial port.  Maybe I could plug a flash drive into the USB port?  Nope... no USB port.

So even though the Mac fired up and ran amazingly well, and even though we found the data, there was no way to get it to a modern computer.  Still, the problem could have been worse.  It was promising that at least the data was in a format that could be read by Excel.  If it was in some old, unsupported database format or something like that, there would have been very little hope.

As it turned out, before I had to reverse engineer AppleTalk, they found another source for the data, and I was off the hook.  Still, this whole process got me thinking.  Digital data seems wonderful for archiving because it doesn't change over time, but really it has some significant problems.  First is the issue that bits are volatile and they can be lost.  We were lucky that this computer booted at all after decades of non-use.  Second is that the breakneck speed of change in the computer industry has led us to changing standards that cause many devices (and their data) to become orphaned.  My Mac SE still runs, but it is an island.  I can't figure out any good way to get it to interact with modern computers (without committing to a lot of work).  We are rapidly speeding away from storing our data on rotating disks of rust and moving to solid state drives.  How long will it be until computers with Ethernet can't be connected because everyone uses wireless?  How long until we stop storing images as .jpg files, or (a more immediate concern) software stops supporting my old DSLR's raw image format?

What will archeologists discover when they dig through our stuff many thousands of years from now?  Will they be able to learn anything about our society?  Will they be able to resurrect YouTube to find out the cat worshiping people that we were?  Or will they be left with unrecognizable petrified hard drives with no explanation of what they meant to us; no idea of the secrets they contained?

I once spoke to Brewster Kahle about this.  His internet archive project is a truly outstanding piece of work, but he knows that even if he could be successful at archiving every piece of information on the network, he still has the problem of changing data formats that may render it all obsolete.  The data we archive needs to be taken care of and kept alive.  Storage isn't enough.

This is why I appreciate it when people resurrect old systems, as with the old version of Mac OS above.  It is a service to us all that they keep history alive.  So while this little 1/3 scale Mac didn't make it possible to retrieve Monika's thesis data, it represents a step toward keeping our digital legacy alive.  To that I say Thank You!

Read more information about the Mini Mac project here.

Monday, August 19, 2013

Hold on to your phone, don't send it to heaven, and why accelerometers don't really measure acceleration.

Here's a little advice.  Hold on to your phone.  There has been some discussion at work recently about an app that encourages you to throw your phone in the air.  When you do, it tells you how high you threw it. That's it.  Of course, it tells you how you did compared to others, thus encouraging you to throw your phone higher and higher until... well... I would image this story rarely ends well.  Eventually I would think either the smartphone isn't caught and it splatters on the concrete, or worse, it hits an unwitting bystander on the head.  The app is called SMTH which stands for "Send Me To Heaven," and it seems to be gaining popularity.  Resist!  I must say that I do love the quote in the marketing of this app, “Probably the last game I´ll ever play on my current phone.”  Someone is having fun.

Still, it is interesting to understand how this application works.  Some of you may assume that the GPS is used to track the altitude of the phone when you throw it, but this is not the case.  GPSes are notoriously bad at calculating altitude, so they cannot be accurate enough to measure a throw of a meter or two.  Most phones have no altimeter.  So how do they measure how high the phone is thrown?  It turns out that they use accelerometers and the same principle that makes astronauts experience weightlessness on the International Space Station, even though they are in orbit and very much under the effects of Earth's gravity (Disclosure: the truth is that I have no information about how they do it, but I know how I did it a few years ago and it is all about accelerometers).   You can learn a little bit about how accelerometers work in the little video I put together for you on part of this subject.

To understand how this is used to calculate height of a thrown smartphone, you'll have to read on...

Saturday, August 10, 2013

The Perseid meteor shower is really just bugs on our windshield

Perseid meteors over the ESO’s Paranal Observatory in Chile - Photo: S. Guisard
Nature is putting on a show and you are invited.  Every year at about this spot in the Earth's orbit, we run into some debris in space. We move through a cloud of rocks and ice (which are remnants left behind comet Swift-Tuttle long ago) much the same way that a car driving through a swarm of bugs on the highway may get a few on its windshield.  Fortunately for us, the splat made by this debris is much more interesting to look at and not something we need to clean off next time we stop at the service station.  Instead, these bits are vaporized as they hit the atmosphere, and they leave behind lovely trails that we call the Perseid meteor shower.  Don't worry, this space stuff is pretty small, so none of it will make it to the surface - there isn't anything to worry about - there is just a lovely view.

To see this light show best, find a dark place, away from city lights.  You don't need a telescope or binoculars.  The best tool for watching this show is just your naked eyes.  Lie on your back and look up (I like imagining that I'm floating in space and the entire Earth is my giant backpack, but that is just me).  It can take about 30 minutes for your eyes to fully adjust to the darkness.  You should start to see shooting stars... as many as 50 per hour.  Go on and give it a try.  The peaks from August 11-13, so don't miss it.  ...and to all you space dust particles from Swift-Tuttle, watch out for that big blue marble headed toward you!

Monday, August 5, 2013

Mini quadcopters to take sensor measurements!

We love the quadcopters here at Digital Diner.  It seems that every day someone is doing something new and cool that involves a quadcopter one way or another.  Today is no different.  A former colleague of mine, Chris Fustings has been having fun with some very tiny little quadcopters.  The Crazyflie is a nano quadcopter from a company called Bitcraze in Sweden.  It is a slight 19 grams in weight and a mere 9 cm span from motor to motor, but still it includes sensors and controls that let it fly like the big boys.  The specs are pretty impressive.  For $120-$180 you get, 7 minutes of flight time, a low energy radio, wireless firmware updating, a 72Mhz 32bit processor, 3-axis accelerometers, gyros, magnetometers and altimeter in a cute little package that can carry a 5-10 gram payload.  Pretty impressive if you ask me.  The 10 degree of freedom (DOF) version of the quadcopter has enough sensors and processing power to do sensor fusion and stabilized flight, although the software has not yet been written (sounds like an opportunity!).

So what can you do with these?  Chris took one and added a distance sensor to it so that he could do a little obstacle avoidance.  His blog does a nice job of laying out the steps so that you can do the same should you desire.  This is a promising start, but he has dreams of doing much more.  He imagines setting up one of these little gadgets to autonomously explore and measure a 3 dimensional grid by autonomously flying to predetermined locations and taking sensor readings.  Quite cool.  Imagine fitting one of these guys out with a geiger counter (a really small one) and flying it around the Fukushima power plant to create a 3D map of radiation.  We just need some really light weight sensors and a really good method for determining 3D location.   You should get to work on this right away.

Take a look at Chris' blog post, get one of these little gadgets and have some fun!  Nice job Chris.