jgarzik
Legendary

Activity: 1596
Merit: 1164
|
 |
September 21, 2013, 02:53:03 PM Last edit: September 21, 2013, 03:14:52 PM by jgarzik |
|
Note that Jeff doesn't know how MasterCoin protocol works, I think he thinks that it is similar to colored coin approach, but it is not!
Let's not put words into others' mouths, shall we? MasterCoin is almost exactly like my original pybond scheme -- which had nothing to do with colored coins whatsoever. Which is, in turn, similar to Mike Hearn's bond proposal from years ago. The only similarities are higher level concepts of "applying a protocol on top of bitcoin transactions." An earlier proposal would have enabled the attachment of mastercoins (or bonds) to each txout: https://github.com/bitcoin/bitcoin/pull/1809 After much discussion in the community, it was decided that OP_RETURN was superior because it provably enables pruning (thus removing UTXO bloat automatically). Been there, done that, all these problems have already been thought through. (The difference is that with colored coins you have to spend a specific UTXO, but with MasterCoin you can spend any coin.)
Just like my pybond design, which had absolutely nothing to do with colored coins. 1. Plain multisig is a viable option. Bloat prevention can be a policy of client: it should spend these multisig outputs either right away, or after some time. 2. P2SH+multisig is possible and easy to implement, but it isn't clear if it is necessary. In any case, it is easy to add support for it. 3. OP_RETURN offers NO advantages, only disadvantages. it shouldn't even be considered.
UTXO bloat is not a policy of the client, if the protocol is specified to store MasterCoin data in the UTXO set. The quoted message is hand-waving away additional burdens placed upon the network -- and every bitcoin user -- by MasterCoin. It is simple, provable engineering fact that storing data in transaction outputs makes block validation, double-spend checks and other critical consensus operations more expensive. More RAM is used on average. In general, it burdens the entire network. UTXO is our most critical resource currently. We have already made one change in recent times to ensure that UTXO growth is more constrained in the future: https://github.com/bitcoin/bitcoin/pull/2577 Big mining pools also have custom patches on top of PR #2577, that ignore transactions they feel bloat the UTXO set. Even more relevant to MasterCoin, at least two pools that I know of elide transactions that appear to transmit data rather than monetary value, trying to prevent another episode of dumping wikileaks cables into the blockchain. Please work with the community, not against it. The blockchain free-rider problem is very real, and deserves thought. (EDIT: well, OP_RETURN has one advantage: it makes Bitcoins more valuable through monetary deflation  Incorrect. All current proposed OP_RETURN schemes set nValue==0, because nobody wants to burn money. Old "sacrifice" schemes have been dropped in favor of miner fees or anyone-can-spend outputs.
|
Jeff Garzik, Bloq CEO, former bitcoin core dev team; opinions are my own. Visit bloq.com / metronome.io Donations / tip jar: 1BrufViLKnSWtuWGkryPsKsxonV2NQ7Tcj
|
|
|
killerstorm
Legendary

Activity: 1022
Merit: 1033
|
 |
September 21, 2013, 03:35:29 PM |
|
Note that Jeff doesn't know how MasterCoin protocol works, I think he thinks that it is similar to colored coin approach, but it is not!
Let's not put words into others' mouths, shall we? MasterCoin is almost exactly like my original pybond scheme No, it doesn't. Again, MasterCoins are associated with addresses, not with transaction outputs. It is a completely different design. MasterCoin user can own MasterCoins even if he doesn't own any UTXOs. It isn't the case with pybond, is it?
|
|
|
|
ASICSRUS
Member


Activity: 70
Merit: 10
Expert Computer Geek
|
 |
September 21, 2013, 03:56:05 PM |
|
Big mining pools also have custom patches on top of PR #2577, that ignore transactions they feel bloat the UTXO set. Even more relevant to MasterCoin, at least two pools that I know of elide transactions that appear to transmit data rather than monetary value, trying to prevent another episode of dumping wikileaks cables into the blockchain. ...
i'd imagine it's possible to filter data type transactions coming from only certain countries that the pool don't like? 
|
|
|
|
cunicula
Legendary

Activity: 1050
Merit: 1003
|
 |
September 22, 2013, 08:46:19 AM |
|
For what it is worth (probably negative to most people), I am pretty sure that killerstorm is right on this.
Mastercoin allows for state dependent transactions. Bitcoin doesn't. This is severely limiting. Almost everything I have ever wanted to see added to bitcoin cannot be added because bitcoin does not allow state dependent txn outcomes (e.g outcome of a txn cannot depend on the realization of a future state variable such as difficulty.)
|
|
|
|
|
|
DobZombie
|
 |
September 22, 2013, 09:03:27 AM |
|
For what it is worth (probably negative to most people), I am pretty sure that killerstorm is right on this.
Mastercoin allows for state dependent transactions. Bitcoin doesn't. This is severely limiting. Almost everything I have ever wanted to see added to bitcoin cannot be added because bitcoin does not allow state dependent txn outcomes (e.g outcome of a txn cannot depend on the realization of a future state variable such as difficulty.)
Can somebody explain 'state dependant?'
|
Tip Me if believe BTC1 will hit $1 Million by 2030 1DobZomBiE2gngvy6zDFKY5b76yvDbqRra
|
|
|
|
zathras
|
 |
September 22, 2013, 09:20:12 AM |
|
I'd like to seek further validation of my transaction engine so I've opened up the block explorer component of Masterchest @ https://masterchest.info. This is very alpha so there may be bugs, errors, mistakes and so on present - feedback is most welcome so please post about any errors or unexpected behaviour. 
|
|
|
|
|
Tachikoma
|
 |
September 22, 2013, 07:02:51 PM |
|
I'm back in action after being tied to the bed for the last week. I've released ' Masteroin-ruby' which is the library behind mastercoin-explorer.com. I am trying to keep the readme up-to-date whenever I update it. The latest version supports sending multisig Mastercoin-transaction via Bitcoin-Qt (or Bitcoind). However as always please be very careful what you do. It's very basic but the same logic could be used by any full-fledged Mastercoin client. Mastercoin-explorer will also recognise multsig transactions now. I've send three (invalid, because the sending address did not buy from Exodus, did not want to send real coins until spec is finalised) multisig transactions, one, two and three. As always I hope all you other developers working around can also try my implementation so we can finalise the spec. A few things I came across. - Sequences for multisig public keys need to start with 01 instead of 00. If you don't do this more often then not the public key won't be a valid ECDSA point.
- My transactions were being flagged as non-valid whenever I used 0.00006 for the multisig output, changing this to 0.00006 * 2 fixed this for some reason.
- Bitcoin-qt doesn't recognise multsig-transactions as outputs ready for spending. You will have to manually redeem those outputs.
Thanks! This should now be fixed. I'd like to seek further validation of my transaction engine so I've opened up the block explorer component of Masterchest @ https://masterchest.info. This is very alpha so there may be bugs, errors, mistakes and so on present - feedback is most welcome so please post about any errors or unexpected behaviour.  Woops; grats! Diversity is a good thing 
|
|
|
|
|
|
|
zathras
|
 |
September 23, 2013, 05:47:43 AM |
|
I'm back in action after being tied to the bed for the last week. I've released ' Masteroin-ruby' which is the library behind mastercoin-explorer.com. I am trying to keep the readme up-to-date whenever I update it. The latest version supports sending multisig Mastercoin-transaction via Bitcoin-Qt (or Bitcoind). However as always please be very careful what you do. It's very basic but the same logic could be used by any full-fledged Mastercoin client. Mastercoin-explorer will also recognise multsig transactions now. I've send three (invalid, because the sending address did not buy from Exodus, did not want to send real coins until spec is finalised) multisig transactions, one, two and three. As always I hope all you other developers working around can also try my implementation so we can finalise the spec. A few things I came across. - Sequences for multisig public keys need to start with 01 instead of 00. If you don't do this more often then not the public key won't be a valid ECDSA point.
- My transactions were being flagged as non-valid whenever I used 0.00006 for the multisig output, changing this to 0.00006 * 2 fixed this for some reason.
- Bitcoin-qt doesn't recognise multsig-transactions as outputs ready for spending. You will have to manually redeem those outputs.
Thanks! This should now be fixed. I'd like to seek further validation of my transaction engine so I've opened up the block explorer component of Masterchest @ https://masterchest.info. This is very alpha so there may be bugs, errors, mistakes and so on present - feedback is most welcome so please post about any errors or unexpected behaviour.  Woops; grats! Diversity is a good thing  Great work Tachikoma, busy getting the wallet ready but I'll take a look at expanding the Masterchest transaction engine to handle multisig too - thanks for all your work 
|
|
|
|
|
superfluouso
|
 |
September 23, 2013, 07:06:03 AM |
|
Thanks for the offer, though as an "investor" this is slightly annoying. Why do you feel you need to separate yourself from the rest of those who may be interested in pursuing the outstanding bounty? As someone presumably qualified, why not jump in with your skills and either A) make bitcoin better by further developing this mastercoin protocol layer or B) make bitcoin better by illuminating the supposed folly of a mastercoin layer (or anything like it in the future) and thereby protect it? If JR and the board want to take you up on your offer, obviously some like myself won't/can't stand in the way, but it appears JR's feelings on this issue were made clear earlier in the thread. I can appreciate your enterprising nature, however, I feel its distracting to the goal at hand. Personally, I'd much rather see talented devs contribute to this project regardless if its favorable or not, just to provably show whether or not this can work and sustain itself. It's not like 40 BTC is going to change your life today anyways...why not just jump in and contribute if you have something to offer?
|
|
|
|
|
Peter Todd
Legendary

Activity: 1134
Merit: 1235
|
 |
September 23, 2013, 07:27:39 AM |
|
It's not like 40 BTC is going to change your life today anyways...why not just jump in and contribute if you have something to offer?
I've got other paid and unpaid work competing for my time.
|
|
|
|
killerstorm
Legendary

Activity: 1022
Merit: 1033
|
 |
September 23, 2013, 07:37:11 AM |
|
Thanks for the offer, though as an "investor" this is slightly annoying. Why do you feel you need to separate yourself from the rest of those who may be interested in pursuing the outstanding bounty? As someone presumably qualified, why not jump in with your skills and either A) make bitcoin better by further developing this mastercoin protocol layer or B) make bitcoin better by illuminating the supposed folly of a mastercoin layer (or anything like it in the future) and thereby protect it? If JR and the board want to take you up on your offer, obviously some like myself won't/can't stand in the way, but it appears JR's feelings on this issue were made clear earlier in the thread. I can appreciate your enterprising nature, however, I feel its distracting to the goal at hand. Personally, I'd much rather see talented devs contribute to this project regardless if its favorable or not, just to provably show whether or not this can work and sustain itself. It's not like 40 BTC is going to change your life today anyways...why not just jump in and contribute if you have something to offer? Your arrogance is sickening. Mastercoin is a commercial endeavor, it got more than $500k in funding. But you're annoyed by the fact that people aren't willing to work for free. WTF is wrong with you? BTW I think accepting retep's offer (or a similar offer) is the only way to save Mastercoin from failure. But you can keep thinking that trusting design to amateurs is a good idea. People like J. R. Willet and ripper234 might be good software developers, but it's quite obvious they didn't have any experience in designing decentralized cryptographic systems and protocols for them. On the other hand, retep is one one few people who know Bitcoin protocol inside-out (not just how to use it, but design considerations which go into it).
|
|
|
|
|
superfluouso
|
 |
September 23, 2013, 07:42:34 AM |
|
It's not like 40 BTC is going to change your life today anyways...why not just jump in and contribute if you have something to offer?
I've got other paid and unpaid work competing for my time. Fair enough..I only hope at some point the potential of MasterCoin will be a bigger lure than a guaranteed 40 BTC to bring you into the mix. 
|
|
|
|
|
|
superfluouso
|
 |
September 23, 2013, 07:50:08 AM |
|
Thanks for the offer, though as an "investor" this is slightly annoying. Why do you feel you need to separate yourself from the rest of those who may be interested in pursuing the outstanding bounty? As someone presumably qualified, why not jump in with your skills and either A) make bitcoin better by further developing this mastercoin protocol layer or B) make bitcoin better by illuminating the supposed folly of a mastercoin layer (or anything like it in the future) and thereby protect it? If JR and the board want to take you up on your offer, obviously some like myself won't/can't stand in the way, but it appears JR's feelings on this issue were made clear earlier in the thread. I can appreciate your enterprising nature, however, I feel its distracting to the goal at hand. Personally, I'd much rather see talented devs contribute to this project regardless if its favorable or not, just to provably show whether or not this can work and sustain itself. It's not like 40 BTC is going to change your life today anyways...why not just jump in and contribute if you have something to offer? Your arrogance is sickening. Mastercoin is a commercial endeavor, it got more than $500k in funding. But you're annoyed by the fact that people aren't willing to work for free. WTF is wrong with you? BTW I think accepting retep's offer (or a similar offer) is the only way to save Mastercoin from failure. But you can keep thinking that trusting design to amateurs is a good idea. People like J. R. Willet and ripper234 might be good software developers, but it's quite obvious they didn't have any experience in designing decentralized cryptographic systems and protocols for them. On the other hand, retep is one one few people who know Bitcoin protocol inside-out (not just how to use it, but design considerations which go into it). I'm afraid you misunderstood - at no point did I imply he should work for free. I am simply questioning why he felt a need to work outside of the already established framework to utilize the 500k in funding. I'm pretty certain if he just did what he proposed, he'd avail himself to a large portion of the current and future amounts made available anyways. Or maybe I just had too many beers. I'll stay quiet..carry on.
|
|
|
|
|
killerstorm
Legendary

Activity: 1022
Merit: 1033
|
 |
September 23, 2013, 07:52:56 AM |
|
Fair enough..I only hope at some point the potential of MasterCoin will be a bigger lure than a guaranteed 40 BTC to bring you into the mix.  People who see the potential in ideas behind Mastercoin are working on similar systems: BitShares, Freimarkets, etc. Why would they want to make a system based on J. R. Willet's inept design if they can design a better system on their own? Really, the only potential advantage of Mastercoin is funding it got. There are people who would like to work in the field of cryptocurrencies, but have to work elsewhere to feed themselves and their family. So actually paying for work can help Mastercoin to attract qualified people... At least in theory.
|
|
|
|
killerstorm
Legendary

Activity: 1022
Merit: 1033
|
 |
September 23, 2013, 08:23:00 AM |
|
I'm afraid you misunderstood - at no point did I imply he should work for free. I am simply questioning why he felt a need to work outside of the already established framework to utilize the 500k in funding. I'm pretty certain if he just did what he proposed, he'd avail himself to a large portion of the current and future amounts made available anyways.
Or maybe I just had too many beers. I'll stay quiet..carry on.
Do you mean that $25k competition? Well, retep's report will likely recommend something "controversial", like a more complex design, keeping data in a separate chain etc. Because it just makes sense, I think all Bitcoin core devs will agree with this. J. R. would not like this, so he will award him no points, and will instead reward people who work on Mastercoin implementation which goes along the lines of Mastercoin spec written by J. R. himself. The thing is, this spec is a bit too half-assed, so while in short term it will give some tools which kinda sorta work, in the long run adherence to overly simple design will cripple Mastercoin. I really don't see why spending ~1% of funding received on a professional analysis is a bad idea, while relying only on people who join the competition is a good idea. I mean, it is theoretically possible that something good will come out of this competition, but why not increase chances of getting something good out of it by spending a bit extra? After all, time is crucial here. I understand that if retep's offers is accepted, that will piss off developers who are already working on Mastercoin (Tachikoma et al.). But they are just not offering same kind of thing as retep can do, so it's fair: different kind of reward for a different kind of work.
|
|
|
|
Peter Todd
Legendary

Activity: 1134
Merit: 1235
|
 |
September 23, 2013, 10:32:31 AM |
|
Well, retep's report will likely recommend something "controversial", like a more complex design, keeping data in a separate chain etc. Because it just makes sense, I think all Bitcoin core devs will agree with this.
FWIW if I think the best design for MasterCoin is to make use of the Bitcoin blockchain that's exactly what I'll recommend MasterCoin do, and I'll explain how to do it in a way that is as difficult as possible for Bitcoin to block. While I am known in the community for my hard-line stance on scalability and decentralization, I don't think we want to kid ourselves - if there are technical means and economic motivations to stuff data into the blockchain we should face that fact and understand the threat now in detail. Fortunately I think those incentives aren't there, or required, but I'd be fooling the Bitcoin community if I didn't consider all the possibilities carefully and in the interests of the MasterCoin community.
|
|
|
|
|
Tachikoma
|
 |
September 23, 2013, 12:50:49 PM |
|
I've added a quick JSON-API on top of mastercoin-explorer. You can find the documentation here.
|
|
|
|
|
mindtomatter
|
 |
September 23, 2013, 04:23:35 PM |
|
Is JR the one assigning the winner of the contest, or is it the board? Seems like it should be the board.
|
|
|
|
dacoinminster (OP)
Legendary

Activity: 1274
Merit: 1050
Rational Exuberance
|
 |
September 23, 2013, 04:28:01 PM |
|
I respect retep's knowledge of bitcoin very much, but as killerstorm suggested, I expect he would recommend a course of action we are unlikely to follow. For instance, his proposal suggests that he may conclude that MasterCoins as a currency may not be needed at all. I personally believe that MasterCoin's biggest competitive advantage is not the funds we have raised, but the simplicity of our protocol. Because of that simplicity we'll have distributed exchange, smart property, and distributed betting far sooner than any of our competitors. If stabilized currencies are possible (I believe they are) we'll do that first too. First means a lot in this game, but in addition to being first, we'll also have tremendous flexibility. Each additional feature we add will take less work than our competitors have to expend with their more complicated solutions. It may be that my naive design has missed something important that will bite us later, but I expect someone would have exposed a problem like that by now. If something does come up, I'll be very surprised if it can't be fixed. I expect that we will be spending a lot of money over the next few months to make sure that progress on this project remains blisteringly fast. 
|
|
|
|
|