[FEELER] Bitcoin bounty project

[JoelKatz]

I know a lot of people have informally offered bounties to get various things bitcoin-related done. What I'd like to do is start a project
specifically to organize bounties. The primary aim would be to make badly needed usability improvements and fix bugs in the bitcoin client.  

I can commit at least some of my time and will put my reputation behind the project. I can provide hosting for the project web pages.

What I'm asking for at this time is primarily a response on whether people are interested, whether my basic plan makes sense, and whether there are any other people who are willing to take on larger roles in the project. Also, if this is duplicative of other efforts, I'd prefer to combine forces than have two projects.

Here's the basic idea:

1) All project records would be completely open. All project funds would be held in accounts with well-known bitcoin addresses. All payouts would be documented: amount, recipient (possibly by psuedonym), and reason.

2) The project would take suggestions for most wanted bounties. The primary goal would be improving the bitcoin client to make the client more useful for ordinary people, especially those unfamiliar with bitcoin internals. But improvements to other bitcoin-related open source projects would be acceptable too. Bounties could be awarded in stages: First to offer patch, first to submit pull request, changes accepted into mainstream client, and so on.)

3) Donations would be primarily through a small number of fixed addresses and separated by purpose for project transparency. Plans for project
termination would be pre-arranged (perhaps donation of funds to a charity chosen from a list, perhaps refunds where possible). Each address would have a focus, for example one address just for client usability bounties, one just for client bugfixes, and so on, so donors could focus their donations on their areas of interest. (Probably no more than 5, for sanity reasons.)

The project would be strictly limited to improving bitcoin-related open source projects with an initial primary focus on improving the usability and
security of the bitcoin client and a secondary focus on other improvements (such as bugfixes) to the bitcoin client.

Hopefully, I could transfer most of the day-to-day control and decision making power over the project to a group of voting members. Sadly, I would have to make myself ineligible for claiming bounties.

The biggest issue for me is the security of holding other people's money. If anyone knows a better way than transferring coins to offline encrypted
wallets with emergency key backups printed on hidden sheets of paper, I'd love to hear it. Perhaps I can find three trusted people and give them each 2/3 of the key as an emergency backup in case I get hit by a bus.

[Serith]

Donated a bit for this project, I wish you best of luck!

[JoelKatz]

I just noticed the other thread on bounties, and it doesn't seem like there are any bitcoin bounty projects that are particularly far along. I'd be happy to consolidate efforts.

[bitstarter]

I just noticed the other thread on bounties, and it doesn't seem like there are any bitcoin bounty projects that are particularly far along. I'd be happy to consolidate efforts.

Hi Joel - yes, you might want to check out this thread:

https://bitcointalk.org/index.php?topic=35905.new;topicseen#new

Maybe, you can find some people there who could help 

[bitplane]

As per the other thread, I have a domain and enthusiasm and a desire to write code.

The project would be strictly limited to improving bitcoin-related open source projects with an initial primary focus on improving the usability and security of the bitcoin client and a secondary focus on other improvements (such as bugfixes) to the bitcoin client.

I'd prefer to have something for all open source projects rather than just Bitcoin, partly for my own selfish needs.

The biggest issue for me is the security of holding other people's money. If anyone knows a better way than transferring coins to offline encrypted wallets with emergency key backups printed on hidden sheets of paper, I'd love to hear it. Perhaps I can find three trusted people and give them each 2/3 of the key as an emergency backup in case I get hit by a bus.

I'd like to have no liability at all, hell, I'd really like it if it were completely decentralized. What do you think about the "many sponsors per project" idea from the other thread? There is the risk of money being lost and/or not paid out if someone loses their password, but it would be risk-free from the site's point of view.

Anyhow, count me in.

[JoelKatz]

I'd prefer to have something for all open source projects rather than just Bitcoin, partly for my own selfish needs.

You know, I was thinking that it should be bitcoin only to promote bitcoin. But then it occurred to me, if bounties are paid in bitcoins, it will serve to get people interested in bitcoin regardless of what the bounty is for. I'm not 100% sure how I feel about this though.

The biggest issue for me is the security of holding other people's money. If anyone knows a better way than transferring coins to offline encrypted wallets with emergency key backups printed on hidden sheets of paper, I'd love to hear it. Perhaps I can find three trusted people and give them each 2/3 of the key as an emergency backup in case I get hit by a bus.

I'd like to have no liability at all, hell, I'd really like it if it were completely decentralized. What do you think about the "many sponsors per project" idea from the other thread? There is the risk of money being lost and/or not paid out if someone loses their password, but it would be risk-free from the site's point of view.

That's another issue that I'm not completely sure how I feel about. I'm a bit concerned that while decentralization will create a lesser risk of a large amount of projects funds being lost or stolen, I think it will create a greater risk of a smaller amount of funds being lost or stolen. I think I'd have to see (or work on) more detailed proposals.

[John Tobey]

I for one would work more on FOSS and Bitcoin if it paid better--more reliably, more directly.

I'm skeptical of the bounty model where the first to finish gets the prize, while the loser may have done better work.  Could you envision a variant where the pot is split according to value contributed after (say) 30 days?  And if value contributed is less than 100% of what was requested, part or all goes back to the donors?

[JoelKatz]

I'm skeptical of the bounty model where the first to finish gets the prize, while the loser may have done better work.  Could you envision a variant where the pot is split according to value contributed after (say) 30 days?  And if value contributed is less than 100% of what was requested, part or all goes back to the donors?

That's another tricky bit. You want to get high quality code, and one good criterion for that is the code getting accepted into mainline. But you don't want to punish someone for something outside of his control. That's another issue that's going to need thinking about.

I was thinking staged bounties. So many bitcoins for the first working implementation. So many for a high-quality implementation and a pull request. So many for getting the pull request into mainline. I'm not sure that's the best approach, but I do prefer to avoid overly-subjective criteria. I could imagine something like 50% earned immediately and 50% distributed after 30 days, including other contributors if appropriate.

[amincd]

As I've mentioned, I think this is a great idea.

Regarding this:

3) Donations would be primarily through a small number of fixed addresses and separated by purpose for project transparency. Plans for project
termination would be pre-arranged (perhaps donation of funds to a charity chosen from a list, perhaps refunds where possible). Each address would have a focus, for example one address just for client usability bounties, one just for client bugfixes, and so on, so donors could focus their donations on their areas of interest. (Probably no more than 5, for sanity reasons.)

This is an excellent way to do it. Having an additional, 'general' address would be good too in my opinon, in case the donor simply wants to donate to any thing bitcoin development related, and give you leeway to decide which area of development to fund.

[indio007]

I don't know if this is possible but it's regards to btc disbursement control. Make it so more than one private key password whatever is needed to access the wallet. 2 keys like they do with nuclear weapons. It could be more than 2. It could be a completely new application, basically multi-peer authentication. Of course people could conspire to thieve but at least a single person having a moment of weakness won't screw it up.

[Sukrim]

That's what he meant with distributing 2/3 of the private key to different people...

To the project itself:
Tahoe LAFS for examle has some 200 BTC in donations that they haven't touched yet, even though they might need some help (if you take a look at their bug tracker) - such a bounty system might help them.

About security:
Highest measure currently seems to be an offline computer, manually transferred blockchain once you want to pay something + manually signed Transaction, which then gets sent to the P2P network from another PC. This however does not make sure that in case somthing unfortunate happens (heart attack because a patch is THAT awesome!) you can still send out bounties.

Maybe you could set up a bounty for that?!

[JoelKatz]

Highest measure currently seems to be an offline computer, manually transferred blockchain once you want to pay something + manually signed Transaction, which then gets sent to the P2P network from another PC. This however does not make sure that in case somthing unfortunate happens (heart attack because a patch is THAT awesome!) you can still send out bounties.

Maybe you could set up a bounty for that?!

I'm thinking hack the client to send the change back to the same address, since we don't want any anonymity. That way, only the private keys to the fixed addresses are ever needed. I just use the 2/3->3 method to protect those.