I'd just like to point out how exciting this actually is, because it seems like the reviewers mentioned it but missed the point.
This isn't some vaporware product that some company is geewizzing us with that will likely never get released. The greatness here is that its completely open. Not "going to be open when we're done with it". You can download the schematic now. You can get a dev kit built by someone else now.
For $99 you can start automating your lavalamps right away. The internet of things starts with the plugs. Tres cool.
How will it help me automate my lavalamps? It plugs into a wall socket, but it isn't a wall socket.
It's no different from a hacked Linksys WRT54G or any hacked Linux powered home NAS, except in shape and openness.
And openness isn't a great distinction compared to a hacked Linux powered device, since their closedness has been worked around.
I don't see it as especially exciting - it has no ethernet-over-mains or wireless or switching or any other tech that makes it inherently geared towards home automation, it's just another linux powered print/disk/ssh/web server, isn't it?
Its open. That's the greatness. The dev kit is a little lacking but already it has GPIO pins if you wanted to do a relay and a usb port for a $20 spark-fun "usb-whacker" if you're too lazy for gpio.
The point is that with the schematic and the backing of Marvell, it will be easy for relatively small players to add cool innovative features with custom builds. Every time I find a place to buy a WRT54G, its different inside and takes a lot of googling and fiddling to get any linux at all on its tiny flash.
The Linksys WRT54G or any hacked Linux powered home NAS are finished products. This is an exciting beginning.
Note: I actually chose lavalamps as an example while looking at the little usb lavalamp on my desk and thinking about writing a tiny web app for this little box so that I could plug it into the usb port and have a "computer free web enabled lava lamp" anywhere in the house. The simple (usb to) serial commands to make the lamp come on and change colors would be super easy with this little device.
Those specs are pretty similar to a hackable wireless router or a Gumstix. There are plenty of slow embedded devices available already; I like this one because it is powerful.
This is a classic case of a potentially disruptive technology, by targeting a different market from PCs.
I can't see Microsoft or Intel being able to compete in this low-power, low-margin space, but as this device gets more powerful, and even cheaper, it could creep up into their spaces.
Ubiquitous computing at the MIT Media Lab -- I saw a great presentation on this a couple years ago by Joe Paradiso. Their idea was also to embed networked computers in power strips as a hub for everything else to communicate through. It's called the Plug on this page.
Unfortunately, my experience seems to be that very few wireless manufacturers are really willing to make their hardware "open." This problem seems to have hit OpenMoko as well; I'm guessing this is why they don't have 3G.
if you'd teleport that thing back two decades you'd be burned at the stake for witchcraft. Simply amazing, that kind of tech in such a package. I'm pretty sure they could make a it lot smaller and cheaper still if the demand is there.
No you wouldn't. Go back two decades nobody will be able to tell that the GbE and USB2 actually work -- or even know what you're talking about if you tell them what those two sockets are for -- so the reaction you'd get would probably be along the lines of "ok... so you have a $100 box which plugs into the wall and lights up a red LED?"
There's a decent chance that the GbE can fall back to 10Mbps ethernet, just as it can certainly fall back to 100Mbps ethernet. The 10Mbps Ethernet standard was released in 1982 (http://www.infoworld.com/article/03/05/09/19FEether.sb2_1.ht...), so I'm going to guess that by 1989 there must have been at least one 10Mbps card.
Edit: It occured to me I should check the timeline on 100Mbps ethernet. According to http://www.broadcom.com/company/timeline/ , the first 100Mbps ethernet product was shipped in 1994, so we're definitely looking at 10Mbps. (Note even if that's off one way or the other by a year or two it doesn't matter for this discussion.)
USB will be a problem, because they'll have to reverse-engineer the protocol, but I'm sure with work they could get it to fall back to a USB1 speed that they could figure out how to work with. The problem there isn't so much the USB as figuring out how to send keyboard events, but even then I bet they could figure it out pretty quickly since I'd lay money the USB keyboard standard is deliberately as close to the old keyboard standards as possible.
However, you are grotesquely underestimating the people of 1989 if you think they're going to be shocked by the idea of a "faster serial port with some defined standards for devices", and you're just factually wrong that they'll be confused by an ethernet port. RS-232 serial ports date back to 1969, and survived into the mid-to-late 1990s, for instance.
This is another one of those "there really isn't anything new in computing under the sun" things. It's hard to name an idea that someone from 1989 wouldn't have heard of or understand, they just lack the hardware to experiment with it for $200 1989 dollars.
Have more respect for your predecessors. They were not stupid. They just had much more expensive cycles.
Thank you, good point. Looking at the protocol on Wikipedia (http://en.wikipedia.org/wiki/Autonegotiation) it looks very reverse-engineerable with not much work, but yeah, it won't work out of the box.
RS-232 serial ports date back to 1969, and survived into the mid-to-late 1990s, for instance.
Did they? Funny, I was using one earlier today...
(And liking it. Nothing like plugging a device in and knowing there are no drivers, no 500Mb software downloads, no autoconfig, no software that has to spend minutes "searching" for the device, no wifi, no passwords, and that everything these days speaks 9600,8,n,1. Just beautiful gets-out-of-my-way simplicity. A joyous technology.).
Sorry to nitpick, but teleporting "works" at space dimension, not time dimension; You would have to use a time machine if you want to go back two decades :)
Although you can bring in the time dimension into teleportation by arguing about the time variation that can result from teleporting, since teleported objects would be traveling at the speed of light.
Don't forget time dilation. If you move one end of the wormhole through space, both ends will be off in time too. It's more noticeable if you move it in relativistic speeds, but anything very fast should do the trick quite well.
I searched all those product sites and couldn't find any detailed technical specs. The question that interests me is how can I run Debian/Ubuntu on it.
Other than that, the idea is awsome. I've been waiting for somethinng like this for years.
Mhhm add another network connection, install Tor , privoxy and perhaps some other network encryption / clean up tools. Config with sensible defaults et voila: nice userfriendly device to overcome data retention.
Very cool. I would love to see this market converge with the Arduino scene to create some very novel uses for interaction. Heck, maybe this is the sort of device that could help launch the "Smart Grid" everyone's getting so worked up about.
This isn't some vaporware product that some company is geewizzing us with that will likely never get released. The greatness here is that its completely open. Not "going to be open when we're done with it". You can download the schematic now. You can get a dev kit built by someone else now.
For $99 you can start automating your lavalamps right away. The internet of things starts with the plugs. Tres cool.