Topic: Roadmap to 1.0? (Read 3587 times)

After 0.9 come 0.10 and 0.11; they're not decimals.

Not hard at all. In fact, I think a few (possibly even most) keyloggers take a screenshot on every mouseclick for exactly this reason. The reason not to implement a randomly-changing on-screen keyboard is that it's damn annoying and doesn't really provide much (if any) improved security, and in fact will probably reduce security substantially as it encourages users to use short passwords, due to how damn annoying it is to click a huge number of buttons when the buttons keep shifting positions randomly every time you click one.

Seriously, security features that are annoying are as bad as no security at all because users will actively try to avoid using them properly: people would rather be insecure than annoyed. They may not admit it, but it's the truth. All security features must be designed with non-annoyingness in mind.

How hard would it be to screen-scrape the password screen?
Sure, it's an extra measure of security, no reason not to implement it ... but it won't protect against capable trojan writers.

In defense of the screen keyboard: Maplestory is played by pretty young kids - if they can do it, surely an un-demented grandmother could.

Thing is, that's not very user-friendly for grandma.

Honestly, the usable-by-grandma goal and the grandma-wont-lose-coins goal are mutually contradicting. I'm beginning to wonder if it's even worth trying to achieve both, and frankly if it's not, I'd vote for choosing the coins-are-hard-to-lose goal. Let someone besides the Satoshi client developers worry about how simple and accessible the technology is for grandma--I'm sure other client developers and companies can handle that just fine.

@Gavin

BTW could the way passwords are punched in on the main client to unlock wallets be changed?

What I am thinking about is an on screen keyboard with these features:
1. Keys are shuffled each press.
2. Screen-keyboard window is placed randomly on screen.
(Saw it back when I played MapleStory - never in my online bank  )

That way future BTC key loggers will be out of luck.

To solve 'security' for non-computer people, you'd likely need non-computer wallets (paper wallets...grandparents are already familiar with important documents, or so I read).

For a "user-friendly-version install":
1) Test the user's printer.
2) Load some bootable image onto a flash drive.
3) Restart and boot into the flash drive.
4) Have this pristine world generate and print paper wallets.
5) After it reboots back into mac/windows, (on the paper wallets can be instructions for making the equivalent "watch-only" wallet, and something like *ONLY LET THOSE YOU TRUST SEE OR COPY THIS PART OF THE DOCUMENT* for the private keys).

Obviously the "spend problem" is more complicated...funds could be spent:
Within tiny-Linux again, or via a version of "offline transactions", or you could make every transaction immediately dump the remaining/unspent BTC into the "next" paper wallet...needs lots of wallets and not as user friendly, but good for large transactions.
OR, maybe they NEVER spend the BTC...since banks/lenders can use the blockchain to see how much BTC is in a wallet, normies can just borrow USD against the paper-"savings" BTC wallet as collateral (what Zuckerberg does with his FB shares...tax free!). For smaller transactions this would be ideal. (In fact, a BTC payable credit/debit card would make a month's worth of transactions into one transaction...[for transactions that don't need to be anonymous]).

...clearly it depends, to some extent, on new BTC innovations.

You could make a 25BTC generate now and see if it crashes the network.... Testing is worth 25BTC, isn't it?
You'll have to wait to December to test that everybody orphans your 50BTC generate hack block though.

and also...

I think we can get there by supporting 2-factor coins (computer+phone both sign transactions), real world identities for merchants (EV SSL cert signs an assertion of ownership of a Bitcoin key), paper backups of wallets and so on. No hardware needed beyond a smart phone.

I was the one who proposed using TPMs and secure CPU extensions for protection of private keys. The technology is sufficiently specialized that I don't anticipate regular end users having access to it any time soon. Merchants and exchange operators, on the other hand, could well benefit, but the expertise needed to produce such a solution is rare.

There was talk about using TPM modules to firewall access to private keys. That would help significantly, but I don't think this has advanced to an actual proposal yet.

I'd be really interested to read any existing discussions about how to reach that level of security.  Does it need specialized hardware or a centralised backup system?  (I know about multisig and how that would help, but it seems as though for that to be effective you would still need a trusted place to stand.)

Heh. That's on my list too:



(From my Using Memes to Explain Bitcoin)

People are working on something better than that which won't compromise a node's ability to function as an autonomous fully validating node but will still make things much faster and use less storage.

More important than supporting header only (SPV) mode, but rather the ability to start as a SPV node and sync-up and transition in the background.  But this isn't yet in the immediate pipeline.

By the way, is there any plan to implement a "header-only" mode on the main client, as Satoshi described it in his white paper?

PS.  sorry I realize it has been discussed already.

Hopefully the blockchain won't eat up that terabyte!    jk

I can't immediately think of anything that should be included in the 1.0 release, but I have a comment..  In the "about" page on the satoshi client, it says "This is expiremental software."  Is that going to stay there forever or will it be removed one day?

Indeed.

It's not that he worries, it's just something that we need to prove as working in production.
There's a difference between testing something in a lab (testnet) and making it work in production ... we need to get this level of confidence.

 

I'm kind of surprised that even Gavin worries a bit about this.

This

Roger that.