Quote from: JorgeStolfi on August 25, 2015, 06:15:20 PM

Quote from: kano on August 25, 2015, 02:31:52 PM

Not sure if you are being disingenuous or you really don't understand. I'll give you the benefit of the doubt and assume your stupid.
A few things: [ ... ] but more importantly, XT wont happen.

It is irrelevant which version of the software is being used (Core, XT, or other). What happens after the first big block is mined depends only on how many miners accept big blocks. ("Small block" = at most 1 MB; "big block" = more than 1 MB but not more than 8 MB. See below for BIP100, BIP102, and other complications.) Note that even the Core version is trivially patched to accept 8 MB blocks.

The chain will (theoretically) split once the first big block B(N) gets mined. There will be a "big-block" branch that grows on top of that block, containg perhaps other big blocks; and a "small-block" branch that starts with an alternative small block C(N) with the same block number and contains only small blocks.

If, at that moment, more than 50% of the miners accept big blocks, the big-block branch (mined by them) will eventually grow faster than the small-block branch (mined by the rest of the miners). The split will be permanent and the big-block branch will grow faster.

If, at that moment, less than 50% of the miners accept big blocks, the small-block branch will eventually grow longer. Then the big-block miners will stop mining their branch, which will be orphaned, and they will start mining the small-block branch too. If someone mines another big block, the chain will split again. This situation will repeat over and over, as long as some miner happens to mine a big block. There will be a single small-block chain with recurrent big-block side branches that are orphaned sooner or later.

Sane bitcoiners should want the first case to happen, and as cleanly as possible: namely, with almost all the miners accepting big blocks. Then if someone mines a big block, the few stubborn miners who refuse to accept big blocks will be left mining a slow chain with a huge and growing backlog. In this case, clients who accept big blocks will be fine, while clients who reject them will not be able to use bitcoin until they upgrade.

Blockstream fans who are not totally stupid will want the second case to happen, but also as cleanly as possible: namely, with almost all miners rejecting big blocks. Then, if someone mines a big block, the few miners who accept it will waste time mining on top of it, only to have that branch orphaned right away. In this case, all clients will be fine, except that clients who accept big blocks will see occasional big orphans.

Bitcoiners who are stupid, or are into financial sado-masochism (like those who are planning to sabotage the blockchain voting), will try to get some intermediate situation when the miners who accept bitcoin are neither the vast majority nor a small minority. In this case, the situation will be highly confusing either to the clients who accept big blocks, or tho those who reject them, or possibly for both.

The competing BIPs with other notions of "big block" will make the situation more confusing. Unless Jeff is into sado-masochism too, he should retract his "compromise" 2 MB proposal. It is obvious by now that the Blockstream devs will never accept any compromise solution.

Weren't you supposed to be some sort of a comp-sci professor? I don't know how to be kind about it, but your students are getting screwed...your whole post is chock-full-o'-fail.

Note that to a satoshi-based cyrpto-currency solution the proper terminology is 'longest valid chain'. Bloat-blocks or any other illegality is simply ignored. It doens't matter if the chain stretches from here to China. What this means is that Core would keep right on chunking away even if it got 1% of the hashing (excepting problems with difficulties and absent attacks.)

As for mining, I supposed and Meni Rosenfeld (who, unlike you, actually does understand Bitcoin and algorithms generally) confirmed that the mining effort will be primarily driven by the market value of the product. I would add that when transaction fees become a factor (hopefully some time before the earth stops cooling on core) they would be added to 'market value.' Thus, the trajectory of any XT-like alt forked off the Bitcoin blockchain would have to command a high market value relative to Bitcoin in order to attract mining. It's actually no different than any other sha256 alt (if there are any which still survive.)

The best hope the XT attackers have would be to subsidize the shit out of XT (and probably attack Bitcoin as well) in order to try to get the flame started on mopping up sha256 hashes. I hope like hell they do because I plan to split my coins and sell XT like hell when the opportunity arises. With luck it IS the govt spooks who are behind XT and Gavin's various malfeasance because they have deep pockets to do this subsidization and it would be an opportunity to get some of my tax dollars back.