There is a new test release of MultiBit at:

github.com

Version 0.5.5beta (Encrypted wallets)

Enhancements:
+ New encrypted wallet format - NOT backwards compatible.
   (Unencrypted wallets are unchanged)
+ Fix for bug where encrypted wallets could be opened as unencrypted
+ Arabic translation 100%
+ Automatic backup of private keys on wallet encryption/ change password/ create new private keys
+ Forward port of everything from version 0.4.13
+ Update to bitcoinj version 0.6


Scan of release checklist

Farting out an idea I might play around with later that might be of interest to you---

The ability to generate a one time use wallet and send an amount to it saved as a file.

The file can be emailed, copied or pasted etc.

One the file is dragged on to the client, the client accepts to privikey, dumps the coins onto your wallet, and the file is now worthless.

Why?

1) It'd be cool
2) You could use it to store coins offline... kind of?
3) Great explaining tool
4) It'd be cool
5) If someone wants to accept bitcoin, but didn't give you an address to send it to, plop the file in an email.

just coming from a local bitcoin meet up. i told about what you were doing and someone came up with the idea of using a custom rom on an old/low end android smartphone - might be even cheaper than the rp.

What you are effectively doing here is mailing a wallet file directly. Because regular mail is not secure you would have to encrypt the private keys ie mail someone an encrypted wallet and tell them via a different channel what the password was.

This leads to trust problems. I can email myself an encrypted wallet backup no problem, as I trust myself, but as soon as you mail a private key to another party there is the 'who spent it ?' problem.

Mailing a private key is like having a joint account with someone.

I presume you mean 0.001 and not 0.0001 in your reply.
The MultiBit default is now 0.001 but there is a minimum of 0.0001.

Could you PM me some example transaction ids of transactions that took 60-90 minutes and some that still had not propagated through the network in a day (and tell me which are which!)

Maybe there is a clue in the transaction structure.

Sounds an interesting idea and an interesting evening ! Breadboarding discretes I am happy doing, but custom roms is a bit beyond my skill level I am afraid. Also, I imagine there are ten times as many 'breadboarders' as 'rom programmers' to get ideas from. Not disagreeing with you mind - it is just the time and effort to learn something and skill up.

There was a guy at the Bitcoin Conference in London discussing using encrypted private keys in the way you describe to store donations actually IN an RFID tag. Cheap Mifare tags hold around 4KB. That is only a few transactions but keys are the order of 32 bytes (unencrypted) so the space saving is a big advantage.

I'd like to add extra ticker to MultiBit - can you give me a push in the right direction where to start?
Thanks

There is a new live release of MultiBit at:

multibit.org


Version 0.4.14

Improvements
+ Arabic 100%
+ Fix for chart balance not matching wallet balance



Scan of release checklist

This release is primarily to get the Arabic translation onto multibit.org.

At the moment key handling is a bit limited in MultiBit, primarily because it does not store all the transactions (or even all the unspent transaction outputs). Of course this has its pros and cons.

I have a little acronym for the main things I want to get into MultiBit before it is 'finished' (whatever that means!). It is:

C - get change spendable. The work on tracking the peers that have seen transactions is the preliminary to this.
C - currency support - type in USD, convert to BTC
E - encrypted wallets in the live code
S - speed - faster sync using bloom filters.
H - hierarchical deterministic wallets, which will also give brain wallet support.

Other than the currency support, most of the rest is code that needs to go into bitcoinj so will be written by various people.

Gary Rowe has been working solidly on MultiBit merchant which is an open source solution that anyone can use to make websites that natively use bitcoin. At some point I want to switch effort onto help build that out (as I have more time than him).