5 ms·
In open source communities, impactful contributions yield influence. Here are the top 100 bitcoin contributors: https://github.com/bitcoin/bitcoin/graphs/contri
by LukeHoersten 11y ago
In open source communities, impactful contributions yield influence. Here are the top 100 bitcoin contributors:
https://github.com/bitcoin/bitcoin/graphs/contributors https://github.com/bitcoin/bitcoin/graphs/contributors
Guys like Wladimir, Pieter, Gavin, Cory Fields, Gregory Maxwell and Luke Jr have a voice because they’ve contributed many thousands of lines of code. (Lines of code are only a proxy for impact).
You may have noticed Mike Hearn isn’t in the top 100 contributors list. He is the primary author of the Java implementation of a bitcoin library:
https://github.com/bitcoinj/bitcoinj https://github.com/bitcoinj/bitcoinj
He started it in 2011, definitely early. But a substantial amount of bitcoin core work had already set the path. There are also similar implementations in many different languages but they are not the primary reference implementation for full nodes.
According to Hearn’s blog post:
“I’ve talked about Bitcoin on Sky TV and BBC News. I have been repeatedly cited by the Economist as a Bitcoin expert and prominent developer. I have explained Bitcoin to the SEC, to bankers and to ordinary people I met at cafes.”
Being cited by journalists is not the same as being a primary contributor.
The disagreement between Hearn and the other developers isn’t about whether to increase capacity, it’s about how. Many of the primary full-node contributors believe a hard-blockchain-fork is a risky approach. Lots of work is being done to explore better options, like segregated witness (http://gavinandresen.ninja/segregated-witness-is-cool http://gavinandresen.ninja/segregated-witness-is-cool).
Mike Hearn tried to (very aggressively) push the idea of increasing the block size with a hard fork. In fact the patch allows node operators to vote and 75% adoption is needed. When it looked like that wasn’t going to pan out, he created Bitcoin XT where “Decisions are made through agreement between Mike and Gavin, with Mike making the final call if a serious dispute were to arise.”
So the claim that Hearn not being able to take over decision making power for the bitcoin community is evidence that bitcoin has failed seems to show something slightly different. It shows that open source software methodology of forking and adoption lets the best implementation win and prevents hostile takeovers.
Mike Hearn is not as impactful to bitcoin development as he or recent news would indicate. Mike leaving the bitcoin community has little impact on the future success or failure of bitcoin.
- itistoday2 11y agoExcellent post. And those who want to understand Mike better should read this: https://lobste.rs/s/82pz7r/curated_links_to_understand_dangers_of_bitcoinxt https://lobste.rs/s/82pz7r/curated_links_to_understand_dange...
- makomk 11y agoTo be clear, Mike Hearn's unilateral fork doesn't switch over to the new rules when 75% adoption is achieved, or even when 75% of mining is done by supporters - it switches when 75% of the last 1000 blocks indicate support, which can quite easily happen with less than 75% support just by random chance. This is an astonishingly aggressive threshold, and if he succeeded it's almost certain that Bitcoin would fork into two different, conflicting versions of the transaction history for a substantial length of time.
- LukeHoersten 11y agoYes good point. Thanks for the clarification.
- gus_massa 11y agoThe random chance is very very low. For simplicity, let's say that 50% of the hashing power agree and 50% of the hashing power disagree. Then the probability of getting 750 blocks is 7*10^-59 ( http://www.wolframalpha.com/input/?i=binomial+distribution+n%3D1000%2C+p%3D0.5%2C+at+least+750+success http://www.wolframalpha.com/input/?i=binomial+distribution+n... ) There are approximately 50000 blocks in a year, so the expected time is a lot of millions of years. This changes when most of the hash power agree. If the 74% agree then the chances are 24% ( http://www.wolframalpha.com/input/?i=binomial+distribution+n%3D1000%2C+p%3D0.74%2C+at+least+750+success http://www.wolframalpha.com/input/?i=binomial+distribution+n... ). The calculation is more complicated because the intervals may overlap, but this would take less than a year. So, when more than 70% of the hash power agree, it's time to upgrade, but 70% is a clear majority. (Just for curiosity: With 75% hash power you get a 52% chance. With 76% hash power you get a 78% chance.)
- badgar 11y agoThank you for taking the time to run the numbers. Your parent poster stuck out as obviously wrong based on high-school math.
- someguydave 11y agoThis calculation doesn't address the arbitrary nature of the threshold.