That sounds very inconvenient for merchants and confusing for customers.
Bitcoin needs to be easier to use. As one of the Winklevoss twins said: people shouldn't have to deal with Bitcoin addresses.
Well actually it would be much simpler if you had an app that let you spend your Starbucks credits when you go to that store (with no need for a Bitcoin address to do so). As a payment system Bitcoin *sucks* as it is much easier to use virtually any other existing payment system (why people have got so enthused about Bitcoin as a payment system is actually beyond me).
If you find Bitcoin too hard to use then you use a fiat exchange to get your credits - so those that don't want to see a Bitcoin address don't need to and actually most likely don't want to use Bitcoin at all (all they want is a coffee).
Also trying to shove every single tx in the world into the blockchain is just plain stupid (and doesn't actually make things easier as it will just overload the network to the point you'd be waiting until your coffee got cold before the payment was accepted).
no one is suggesting we shove every single tx in the world into the blockchain, just bitcoin tx should appear on the blockchain.
Actually, that's part of the issue that I have with this discussion. There is this tendency towards tunnel vision
about what scaling means. The way people are approaching this is, "
no matter what transactions comprise blocks at present, we unequivocally need to increase capacity (as much as possible) to include any/all transactions." That's an irresponsible approach, and why we saw the community rally around such a reckless implementation as BIP 101 / XT.
I am not interested in the bitcoin blockchain including any/all transactions. This should be a legitimate topic of discussion, not to be brushed aside. If we want to be technical, an altcoin like Devcoin could be used to make much, much smaller microtransactions than bitcoin. But it doesn't logically follow that we should alter the protocol to include divisibility past the satoshi unit or lower the dust threshold considerably lower than 546 satoshis simply to include these useless transactions. Clearly,
there must be some discussion of what transactions belong on the blockchain, to optimize efficiency and encourage decentralization. what does scaling mean to you? 7TPS with high fees, and with lighting network and or side-chains?
"what transactions belong on the blockchain, to optimize efficiency and encourage decentralization. "
as many as we can get, until there are so many that a typical home computer can handle the traffic and this starts to affect decentralization in a big way. at which point we cap the blocksize and let fees run up, until a better faster internet is available to all. and again we let in as many TX as we can get, until there are so many that a typical home computer can't handle the traffic and this starts to affect decentralization in a big way.
that would be my answer.
we could limit ourselves so that bitcoin full nodes can run in an embedded device, but why would we? limiting ourselves to a typical home computer makes more sense.
thing is you could say bitcoin needs to consume no more then 10% of a typical home computers resources, and you'd still be able to push the TPS to 400. so i find it SILLY to think 1MB is the good limit.
