Oh, sorry, I thought you were a little more familiar with Linux.  My mistake.  

What I wrote above is a series of commands that you put into a terminal window.  To open one, hit the start button, type 'terminal', then click the terminal app.  You should get a text window.  This is similar to a cmd window on MS Windows.

Then you can start typing the commands above.  Here's what they do:

Download the file:

Now it's in your home directory.  A .tar.gz is like a zip file.  To extract it we use the tar program:

Now your home directory has the tar file plus the bitcoin directory.  To see everything there, type 'ls', which is like 'dir' in a cmd window.

Next you go into the directory where the program is.  'cd' means change directory.  The program is three levels in, so:

Above I just shortcutted you in like this:

Either way works.  Then you have to run the program.  Unlike windows, the current directory isn't in the PATH on Linux, so if you want to run a program you have to prepend './' .  Thus:

And the program should start.


If you prefer to work with the GUI, you can do it that way too.  First use Firefox to download the .tar.gz .  It will save it to your Downloads folder.  To get there, click the file folder icon on the left.  Click the bookmark for Downloads.  You'll see the .tar.gz.  You can drag it to your desktop if you'd like.  Right click and choose "Extract here".  You'll see the new folder appear.  Double click it, then 'bin', '32', 'bitcoin-qt'.

wget http://softlayer.dl.sourceforge.net/project/bitcoin/Bitcoin/bitcoin-0.6.2/bitcoin-0.6.2-linux.tar.gz
tar -xf bitcoin-0.6.2-linux.tar.gz
cd bitcoin-0.6.2-linux/bin/32
./bitcoin-qt

Why not add an RTC?  http://datasheets.maxim-ic.com/en/ds/DS3234.pdf

Here's a breakout board for your prototyping pleasure: http://www.sparkfun.com/products/10160

Or this: http://www.sparkfun.com/products/99

Start with Memtest86+.  Run it a few hours and see what you get.  If it always fails in the same spot, you have bad RAM.  If it fails on the same columns at random rows, you probably have incorrect settings.

Corsair Vengeance is overclocker RAM, designed to be run at nonstandard voltages and speeds.  Look up the exact model on Corsair's site and configure their recommended settings (especially voltage - it probably needs to be overvolted) into the BIOS.

apt-get install p7zip-full
7z x whatever.cab

You can also download an up-to-date blockchain here: http://eu1.bitcoincharts.com/blockchain/

I edited what I said a bit to make it more accurate.  It's not a simple majority to fork; it's you have to convince the world that your new system is the new chain is BitCoin and that the old system should be ignored.

Technically, yes: if the change is made and enough people support it, it will be so (or at the very least it would create a permanent fork).

Politically, no: such a change would undermine all confidence in the currency; such a move is unlikely to achieve the necessary support.

The market will set a price, but if I want to hold some savings long-term, I want to know if someone can suddenly resurrect a large number of very old coins and devalue my holdings.  It's not a problem now with most coins in active circulation, but what happens if we end up with a large majority of stale coins?  It's not good for confidence.

OK, I'll play devil's advocate. 

1 can be solved by making people roll it over to a new set of coins every few years.  2 is true and it's why there's no coming 'doomsday'.  I do think there's a benefit though: not being able to know how much currency is in circulation isn't a good thing when you're relying on the rest of it to raise in value.  There's no need to return the coins to circulation, but voiding them after some time has advantages - it would let you know how large the money supply is, and it would allow old blockchain history to be dropped completely after the coins expire.  (The current pruning scheme often can't remove old blocks when some outputs are still unspent.)

This change shouldn't be done in Bitcoin since it would revoke the existing system where coins are good for long term storage.  Who knows how many are on paper in safe deposit boxes with no way to contact the holders to let them know they have to do something to keep them?  But it's an interesting idea for an alt coin.

The weakest link for brute force of any address is the RIPEMD-160 hash.  Yes, you can steal the coins from an address with a different key pair if you happen to find one that has the same RIPEMD-160(SHA256(pubkey)).  Good luck. 

Once coins have been sent FROM the address the pubkey is known and ECDSA is probably the weakest link (about 128 bits of security equivalent).  This may be plausible to attack in the distant future, but probably not in my lifetime unless a cryptographic flaw is discovered.



Pubkey hashes always begin with 1.  It's base58 - I, l, 0, and O are excluded.  The actual address size is therefore 58^33.  That corresponds to 2^(160 (hash) + 32 (checksum)) plus an extra bit or two.

I am an expert.  Yes, it's theoretically possible, but not plausible.  2^160 is a very large number and human civilization will end before a duplicate key is found either accidentally (simply not going to happen) or on purpose (physics imposes some hard limits on computation; we're nowhere near it now, but it will put a halt to moore's law before it becomes a plausible problem, let alone one that's economically worth pursuing).

Edit: I should add, "on purpose" only includes brute force approaches.  Quantum computers can reduce the complexity to 2^80; it's not certain that quantum computers capable of handling this much data are possible, let alone fast enough to process that many operations economically; but it's a possibility.  It's also possible that a cryptographic flaw will be discovered in RIPEMD, SHA, or ECDSA.  For now brute force is the only possible approach, and it's not a plausible problem.

Then you just spend the original inputs again. 

I can't imagine ANY store is going to deliver goods before the tx is at least broadcast, unless it's a regular customer they know and can trust.  For anyone else, they'd have to be online for any transaction.

Still, this is a good point: even if the store broadcasts it immediately your change won't be confirmed for an hour.

$40-60 for something the size of a Kindle.  It's hard to say what a tiny one would cost since there's not much of a market for it.

LCD backlights consume a ton of power.  Unlighted LCDs consume barely any, and regardless this is a device you'd only power up for a few seconds at a time.  eink's nice, but it costs a lot more than a $2-5 LCD.

Here you go:

http://www.mouser.com/Embedded-Solutions/Engineering-Tools/Embedded-Processor-Development-Tools/Development-Boards-Kits-ARM/_/N-8x0x4/


Here's a nice cheap one. The chip is a Cortex M4 with 192KB RAM, 1MB flash, ethernet, USB, LCD drivers, SD card support, and more; the board has some accelerometers, buttons, LEDs, a USB port, and some prototyping leads, all for $15:

http://www.mouser.com/ProductDetail/STMicroelectronics/STM32F4DISCOVERY/?qs=J2qbEwLrpCFMptdjNAVzZeZDfltJ6JKw1GLhrq7db5E%3d

A few thoughts:

The store does not have to send the change.  The device can create a transaction so 0.05BTC goes to the store and .05BTC goes back to you.  There is no need to trust the store to return it or to verify the change.

There is no need to split the inputs into 0.1BTC amounts.  It's just as easy (actually easier) to have a single 10BTC input and send a 3.55BTC output to the store and the remainder as change to yourself.

A 32-bit ARM MCU with 256KB of RAM is only about $10 in single units or $5 in volume.  That plus an SD card to store the blockchain would give you a full-function device.  A CR123A battery would run it for two days of continuous 150MHz operation, and essentially unlimited sleep time.  That's certainly heavier than an 8-bit micro running on a couple of watch batteries, but it's something to consider.

No - the patch is really meant for people who want to control which addresses are associated for anonymity reasons.  It still sends change to new addresses.

You can either sweep the change back to your named addresses, or you might want to look at Electrum.

Nope, it also extends the bitcoind RPC API:

https://bitcointalk.org/index.php?topic=24784.msg446276#msg446276

Yes, the coderr fork (the one I linked) does it.  The patch will probably eventually get merged into the standard client.  Electrum can also do it.