Funny to see it at the top of HN 7 years after I made it! My kids still play with it from times to times (one of the game that lasted the longest in the family)
I used to write agent-based simulations for science education, and one of the main libraries was AgentScript, which I believe is now SimTable [1]. They had developed their own sand table with an agent-based model running on top, and used it to create wildfire models to help plan and train firefighters. I think they also would do modeling of defense scenarios. It was pretty interesting stuff.
I'm sorry, I haven't heard of sand tables before; the agent-based sim concept with lots of people standing around a projector is cool; but is there something better about it being projected on a lumpy surface?
It's a good question, and I honestly think that SimTable has struggled with that.
From a playing perspective, or educational perspective, the sand is great. What happens with wildfires if we make the mountain this big? What happens to the water table if erosion cuts a valley through here? The immediate feedback that you can get by just shaping the landscape differently with your hands is really cool.
From a wildfire training perspective, or other disaster management, it's different because you're dealing with an existing landscape. I've seen them start by projecting a map, then forming the sand with their hands to match the map, and then starting the simulation. But at that point is the 3D landscape needed? You could also project onto a flat table.
I guess the reason is that it helps the fire chiefs or whoever really envision the landscape. "Ah, yes, this is Devil's Pass, I see how the wind is going to blow right through there and blow the fire that way. What if we cut a fire lane here? Ok, that slows it down."
But it could also be a cool technology being used as a solution to a non-problem.
> I guess the reason is that it helps the fire chiefs or whoever really envision the landscape. "Ah, yes, this is Devil's Pass, I see how the wind is going to blow right through there and blow the fire that way. What if we cut a fire lane here? Ok, that slows it down."
Yeah, if you're familiar with topographical maps (the way any decent hiker or fire chief would be), you really don't need the physical 3d layout to envision the landscape.
ah yeah, I was just thinking about that firefighter demo and couldn't understand how the topography was correlated with the surface. (It seems like it'd be super cool if you could scoop out lakes and stuff... I've never seen or played with something like that.)
The tactile aspect of the sand is neat. I can see it as a good learning tool as you described for thinking about wind patterns and road cuts.
The UC Davis AR Sandbox is awesome, happy to see there is still work on keeping the idea alive. I've built one for my school quite some time ago and honestly it's just so cool playing with it I almost wish I had one for myself.
This one specifically seems so much more polished than the UC Davis research project with a dashboard for configuration such as color maps, contour lines, sea level, etc. which had to be changed in config files previously. The calibration also seems automated? I distinctly remember the calibration steps to be a major pain.
Also I'm really thankful for the Sandbox project since it was what first got me into using Linux as my main operation system, since I was the only one in the whole place which any idea what a "Linux" or a "package manager" was.
The only part of this project that still gets me a bit mad is the need for the Kinect, it's been getting harder to find X360 versions of these recently and specially the cable to connect it to a computer (in case of using the XONE version). I really wish there was some kind of hardware alternative to it.
I've made one of these as well using the 'original' setup. We used rags and blocks as I figured a sandbox in the living room matched with two small kids would have been a relative nightmare.
Holding your hand over spots, this making it rain, and having the water "pool" and run as it would is absolutely beyond fun. Move the mountains, and downward it continues to flow.
With a budget projector ($82!), A Kinect (I forget version), and an old graphics card it worked pretty flawless after setup -- probably the hardest part was aligning the sensor with the projector and getting it all actually calibrated -- the next one we build I'll def try this version and note the differences. Calibration with the other method used a cdrom as it's target that you placed around, a bit fiddly.
I'm always a bit pleasantly surprised when following technical terms with a metaphorical meaning (an AR sandboxed playground) and it turns out to be they're used in the literal sense - it is an AR box, filled with sand!
I made one in the past, we’re often asked by schools to make one for them. The limiting factor is the Xbox Kinnects, and needing to solder on a regular USB connector.
Beyond hardware, the other issue is that it is quite an expensive investment for something that lacks direct curriculum/syllabus alignment.
I got to play around with one of these during my stint as a solder monkey for the Tangible Media Group at the MIT Media Lab back in the early 00s. Lots of fun, worth trying out. It was, of course, way more expensive than a setup in the early 20s is.
The Museum of Science and Technology in Syracuse, NY has one of these - despite moving out of Upstate NY a while ago and living in other cities with much bigger, higher budget science museums, it's still one of the most memorable and interesting exhibits I've ever seen at one.
They also had another interesting project using a kinect-enabled display, where you could stand in front of a screen showing a Jurassic scene and throw dodgeballs at velociraptors. The original plan was actually to sync the experience to another display in a museum in Boston, so you'd be playing raptor dodgeball with the other exhibit's patrons over state lines - unfortunately, the other museum never set their half up.
I think that speaks to the question commenters are raising here about why these aren't more common or commercially available. A lot of cool stuff like this is tied to the Kinect and comes with non-trivial setup and maintenance cost.
Cool project, but I don't want kids playing with sand inside.. something like this that uses clay or even Legos would be cool. It couldn't be exactly the same, I understand, since the shadows would ruin the immersion. A whole room version would be cool too that takes into account toys and clothes. Bonus points if it can help encourage tidying up as part of play.
Actual magic/kinetic/hydrophobic sand (what this is intended to be used with, as you can shape it in shapes other than "pile") is pretty self contained, make a reasonably tall tub to play in (maybe 3~5 times as tall as the "sea level") and you should be fine, just make a rule of "no throwing sand meteors" and you should be fine, neatness-wise.
If you put any sort of fine-particles in a inside environment without having a fully enclosed box and anti-containment compartment with double-doors, you are guaranteed to end up with some of those particles reaching every single part of the house.
Once we took apart a confetti+glitter "gun/bomb" on the kitchen counter with strict measures of washing your hand after touching the confetti, don't let it leave the counter and so on, and we still find traces of it in other parts of the house sometimes. Never again.
Well, it's sand anyways. Glitter is annoying because it's highly visible so you're constantly reminded of it, but sand is just like any sort of dust. It's also a bunch of very thin and tiny sheets, so it adheres to surfaces formidably.
Of course, you do your own "annoyance assessment", but a non-perishable, hydrophobic, heavy and visually unremarkable powder has a pretty good value in terms of unit of fun per unit of annoying in my book. Tidier than giving a kid a cookie.
That _is_ a very interesting idea. If every toy had a "home," then it could be a game; every object not in an assigned place would turn red, and you have to find its home.
I have a number of similar AR ideas like this, but thinking about cobbling together an optically aligned projector/depth sensor/camera is the hurdle that I need to get over first. Plus it would be a messy setup to actually deploy.
I'd really love to see some company develop a single unit that had maybe 5000-10000 lumens in a box with an HD camera.
