You aren't a coder, are you? If so, I envision many off-by-one errors in your work.

Cheap is nice. Short term memory -- too close to home for anyone over 40, but one day isn't bad. How fast is it to access, and how large is it relative to Si storage? Those may both be answered in TFA, but I'm too lazy to check. +1 Honesty?

If this is as fast as traditional large storage formats, and it doesn't take considerably more space, it could be interesting to see this applied to swap space or /tmp type storage... especially if the 1 day reliability can be extended through a refresh cycle.

The Myth Busters episode regarding this is several years old. Fingerprint scanning technology has improved rapidly in the mean time.

Like everything, it's hackable if you know the full details, but it gets harder each year, and even when Myth Busters used photo copies and copper etchings to get past them, there were commercially available scanners that these techniques could not fool. I have no reason to believe Myth Busters was aware of the bleeding edge technology, but I do know the technology existed then and is improved today.

"Funny" is close to the correct moderation category for this, but it would be better if I could mod it +1 Sad but True.

The problem with transparent relocation like this is arbitration of which physical source to use when, for example, the local server has a valid "tech" component, and "tech.shashdot.org" also exists.

Worse... on my domain, *.mydomain.tld" resolves to www.mydomain.tld -- it seems like such an implementation could cause my site to fail.

No doubt, many open source projects compile with many warnings. Not that this is an excuse, but open source projects do have a bit of an uphill battle on this front, as there are many compilers out there, and unless they have a build farm to check against them all with, they're not likely to fix everything. [Again, not to make an excuse or minimize your statement -- most warnings I've seen in open source code are easy to fix and not compiler dependent.]

I would imagine its fair to guess that Microsoft builds their OS using the same compiler that Visual Studio uses. Ever work in that environment? Perhaps there are ways to elicit better warning output, but the default (at least when using the UI) is that many warnings that GCC (and other similar class compilers) output, Visual Studio lets go without a peep. That can make code compile warning-free, but in no way is the number of warnings a promise of comparability to quality of what gets produced, eh?

The uneducated are indeed the target market for schools which lack acceptable accreditation. The definition of educated vs. uneducated is not the same as "went to school" vs. "didn't go to school". You should apply a little insight of your own before choosing to get snarky.

I think a plausible possibility is some app you installed did provide these schools with your contact data, but not directly. Instead, through some affiliate website where the site owner gets paid as if you had gone to that site and provided the information/sales call request yourself.

Yes, schools like this prey on the uneducated and exist only to relieve them of their money, but I wouldn't be too quick to assume the schools know that your information was not provided to them legitimately -- especially if it's more than one place that contacted you like this.

Apple provided the API on the iPhone that allows their app to collect the data without any direct decision from the phone's user. It seems to me that "Apple gives them the information" applies, and will continue to apply until users must take specific action (beyond simply installing the app) to allow the data to leave their device.

My question on Android is : how can google not release it under GPL as they are using a linux kernel ?

The parts of their system that are linked against other GPL modules -are- GPL. There are other parts (such as Google Maps, etc.) which are not. GPL is not the virus M$ would have you believe it is; it's very easy to create a system that has plenty of GPL along side plenty of other licenses, all coexisting peacefully.

I suspect you mean http://dictionary.reference.com/browse/moot and not http://dictionary.reference.com/browse/mute.

But I could be wrong, I am on a regular basis.

According to the manufacturer, http://www.embeddedarm.com/products/board-detail.php?product=TS-7800 boots to Linux in 0.69 second. It's a 500MHz ARM-9 based system. I haven't used this board, but I've used others from the same manufacturer; the Linux they provide is Debian-based on the boards I've used.

Mod parent up +insightful.

I've been programming for over 25 years, most of that at the firmware/driver level in C. I went on a hardware quest (for my own edification) about a year ago, and I did an independent study course for credit at my local university. My focus was on Verilog (because that's what's used within my company), and it was truly a different beast than programming! I definitely had to train myself to think differently to accomplish that.

There are many good boards suggested by others in this thread. I used three boards myself. The first board I used was http://www.xgamestation.com/view_product.php?id=40 - containing a Lattice Semiconductor CPLD. The board was a good starter because it was designed for education, and included a lab manual of quite a few projects. Potential downside: the lab manual and tools it comes with are not based on either Verilog or VHDL, but rather ABEL. I did everything in Verilog instead, and still used the lab manual as a source of exercises. The Mach64 board is fairly inexpensive.


I also used a board from http://www.opalkelly.com/ -- this product is a little pricey as an educational device. Their hardware contains Xilinx parts. What I liked most about it though is its USB interface -- first, you can use their tool to download your hardware description without being required to flash the board (where many other boards require you to flash the image to exercise it). Second, and more interesting to me -- Opal Kelly's boards include virtual device libraries, letting you attach hardware to USB-carried channels for data I/O and triggering to a remote application. They include a library to allow you to write a remote application to support the other end of this. I don't know if they support Linux though -- everything I did on the PC end was under Windows.


Finally, a reasonably priced board is http://www.embeddedarm.com/products/board-detail.php?product=TS-7300. This board is not centered around its FPGA, instead it's an ARM9 based CPU board. The board includes full embedded Linux support and all the typical devices you would expect to find, and it has a user-accessible Altera FPGA on board. Their linux distribution includes a tool to send a FPGA bitstream over to the device, letting you quickly test a new hardware description (like the Opal Kelly board, without programming flash). Potential downside here... the manufacturer includes a bitstream image that is used by default, which enables (among other things) its VGA port and a second Ethernet interface - and their open source Verilog for this device doesn't include those, so you sacrifice some of the board to define your own hardware. You don't sacrifice anything you can't use the board without though (there's another Ethernet port; there's a few serial ports, etc., which don't require the FPGA).


Good luck!

I have nothing to add, but how could I possibly not reply to this.

Thank you for honoring me so!

About three years ago I was in Korea, contracting at a mobile phone manufacturer for a while. Their latest phone (Pantech... I can't recall the model number but it was destined for the Russian market) had an accelerometer and they had experimental software to allow you to dial by "Drawing" your numbers in the air. It definitely looked silly to see someone using this, as you had to draw fairly large numbers to make it work, but it worked.