[2015-06-17] Bitcoin update for the next wave of BTC users (by Gavin)

[LiteCoinGuy]

"So...

... maybe there's a misunderstanding on what I'm actually working on. I'm coding a hard-fork that will only fork:

    -After March 1, 2016
    -If 75% of hashing power is producing up-version blocks
    -AND after some brave miner decides to actually produce a >1MB block.

It will be very difficult for anybody to lose money. Even if we assume that some stubborn minority of miners decides not to upgrade (in spite of Bitcoin Core warning them first that the chain consists mostly of blocks with up-version version number, and then warning them that there is an alternative invalid higher-work chain), those miners will be the only people who will lose money. Ordinary, SVP-using users will follow the longest chain, and since at least 75% of hashing power will be on the bigger-block chain, there is no chance of them losing money. The big-block-rejecting-chain will very, very quickly be left behind and ignored.

There will NOT be two active chains, that is just FUD. Anybody running old code will have to willfully ignore the warnings to upgrade to stay on the old chain, and the incentives are so strong to follow the majority I can't imagine the 1MB chain persisting for any significant length of time.

--------

I'm happy to tweak the parameters. The code I'm writing regression tests for right now is:

    -8MB max block size (chinese miners were unhappy with 20 for not-entirely-clear reasons)
    -Earliest fork date 11 Jan 2016 (miners and others want sooner rather than later)
    -Activation when 750 of last 1,000 blocks are up-version (version 0x20000004 to be compatible with sipa's new versioning bits scheme)
    -2 week 'grace period' after 75% threshold reached before first >1MB block allowed to be mined
    -8MB cap doubles every two years (so 16MB in 2018, etc: unless there is a soft fork before then because 16MB is too much)

The code for all that is starting to get right on the edge of "too complicated for consensus," but the individual pieces are all straightforward. I'll write a BIP when I'm done with the code, and, as I've said repeatedly, I'm not stuck to any particular set of parameters."


https://www.reddit.com/r/Bitcoin/comments/39ziy6/eli5_what_will_happen_if_there_is_a_hard_fork/cs7xe9o

[Pro_Crypto_Marty]

This all seems very reasonable, although I'm a perpetual newb when it comes to tech and coding so all I can do is take his word for it.

I still don't quite see why there's so much opposition to the block size increase.

[qwk]

This all seems very reasonable, although I'm a perpetual newb when it comes to tech and coding so all I can do is take his word for it.

I still don't quite see why there's so much opposition to the block size increase.

In simple words, a hardfork is risky. It could go wrong. It's not very likely and the measures Gavin proposed probably are good enough to prevent it from going terribly wrong. Then again, we can't be 100% sure.

[Westin Landon Cox]

75%? At least that's better than Mike Hearn's requirement of an "economic majority" (however that's measured). 75% is significantly lower than the 90% figure others have been repeating in threads on the forum.

If 25% of miners stay on the original chain, then it will adjust to the new difficulty within a couple of months after the fork. Even sooner if other miners recognize that the original chain is continuing.

I'm skeptical that 75% of miners will switch to XT though. If they can't get 75%, what's more likely: those who want the fork will give up or those who want the fork will lower the 75% threshold?

[dserrano5]

75%? At least that's better than Mike Hearn's requirement of an "economic majority" (however that's measured). 75% is significantly lower than the 90% figure others have been repeating in threads on the forum.

75% is activation (allowing for that 2-week grace period), 90% is enforcement.

[Westin Landon Cox]

75%? At least that's better than Mike Hearn's requirement of an "economic majority" (however that's measured). 75% is significantly lower than the 90% figure others have been repeating in threads on the forum.

75% is activation (allowing for that 2-week grace period), 90% is enforcement.

Can you explain what you mean by "90% is enforcement"? Reading what Gavin wrote, it seems clear that 2 weeks after 75% of the hashing power is reached, a miner can mine a >1MB block and it will be accepted by XT. At that point 100% of miners will either be building on the XT-chain or the Core-chain, with no miners building on both.

nullc argues on that reddit thread 75% is too low. The first attempts to mine a >1MB could get orphaned by the Core-chain a few times.

It may take several tries for the actual hashpower majority to decisively overtake (e.g. 25% hashpower gets 3 ahead with about 20% success rate).

A 75% hashpower threshold guarantees there will be competing chains, how long they'll last is less clear while they exist ample opportunity for theft exists; to see you state otherwise is just astonishing to me. To avoid these risks even for soft-forks which of already not is standard changes we already don't enforce to much higher criteria, even though in that case its likely that no divergence will ever begin at all, simply because the risk created if one did happen to form.

[dserrano5]

75%? At least that's better than Mike Hearn's requirement of an "economic majority" (however that's measured). 75% is significantly lower than the 90% figure others have been repeating in threads on the forum.

75% is activation (allowing for that 2-week grace period), 90% is enforcement.

Can you explain what you mean by "90% is enforcement"?

While preparing this answer, I realised there's no difference in this case.

BIP34 and 66 were rolled out using these two percentage thresholds (using actually 95%, not 90). When reaching the first of them, the new rules become valid, but the old rules still apply. When reaching the second level, the old rules stop being valid.

In this case, it makes no sense that the old rules (limit 1 Mb) stop being valid, so effectively there's only one single threshold of adoption.

[Westin Landon Cox]

Thanks. I'm reading about BIPS 34 and 66 now. It looks like 34 was implemented via the procedure you described (75% then 95%). If I understand correctly, 66 will be too, but it's still coming.

[dserrano5]

Thanks. I'm reading about BIPS 34 and 66 now. It looks like 34 was implemented via the procedure you described (75% then 95%). If I understand correctly, 66 will be too, but it's still coming.

BIP66 is between the two levels right now. You can follow it in: http://bitcoin.sipa.be/ver-50k.png