Bitcoin-XT is not a re-implementation, it's a patch on top of Bitcoin Core that adds two features, double-spend relaying (which can be flagged for inspection) and the BIP64 getutxos message (which he needs for Lighthouse). The reason XT exists is because Mike Hearn can't get BIP64 implementation merged into Core (
https://github.com/bitcoin/bitcoin/pull/4351). Gavin happened to be for putting BIP 64 in, so moving forward if this happened, Bitcoin-XT would be a proper fork based on the actual Core codebase, simply with these two differences plus the blocksize changes. The fact that moving to XT is considered a viable option might suggest that there is a much larger "turf war" going on here since at least the middle of 2014.
Thanks!
Indeed there seems to be more going on than a simple technical disagreement.
I won t say that Gavin's proposal is the best one, but at least I understand his position: "with 1 MB blocks the network is close to saturation and will not handle the volume that we would like to see, so let's make the blocks bigger".
The thinking of his opponents (who include Peter Todd, Gregg Maxwell, and Luke Dash Jr.) seems less clear: they say that they are worried about the consequences of bigger blocks, but they have no alternative proposal to deal with impending congestion, and seem to
want to see the network saturate.
Someone on reddit pointed out that most or all of those big opponents work for Blockstream, the company that was supposed to develop sidechains and is now working on a thing called the Lightning Network, Those are projects that would provide fast bitcoin transactions and other bitcoin services (such as micropayment channels) outside the blockchain. That could be a reason for wanting the network to saturate. However, Greg says that they were opposed to big blocks well before creating Blockstram.
Someone else suggested that they may want to see the network saturate so that big non-payment users like NASDAQ and Factom are forced to pay huge fees, like $50 on a 1000 satoshi transaction. That would push common users out of the system and turn bitcoin into a tool for big corporations only. (Peter Todd does not miss a chance to mention that he is "talking to Big Banks".) Sounds like the FUD that only JorgeStolfi would write.
Anyway, last Friday night a handful of reddit users set out, without much planning, to try to saturate the network with small transactions. In the course of 2 hours (23:00 to 01:00 UTC) they put out maybe 30'000 transaction requests. The queues at the nodes got over the 20'000 mark and the backlog took 8 hours to clear. It is not clear what fees the guys paid, but they did not put much money into it. Yet Luke Jr claims that the test only showed that 1 MB is fine. Boh.
Perhaps the real fear is how the system will react to an
intentional hard fork, that is not mandated by a bug...
