A new genesis block isn't just for Solidcoin... as Bitcoin has the same ailment

[2112]

I kinda don't care about the specific algorithm to prune the block chain. But the block marking the pruning need to be called "jubilee blocks" or "Leviticus blocks", in the same way that the origin block is called "Genesis block".

http://en.wikipedia.org/wiki/Jubilee_(biblical)

But seriously:

The only participants of the bitcoin network that require trust are the exchangers. They will always have a cash flow to support the ancillary requirements of the network. The sooner you guys stop thinking about miners (and mining pools) as a centerpiece of the network, the quicker you'll arrive at a workable solution to this problem. Miners (of fresh bitcoins) need to be relegated to the status of peons. The actual value added is in the refining (collecting transaction fees) and the currency exchanging. Only exchangers have the incentive to maintain long-term business presence in the ecosystems. Pretty much everyone else is better served by being fly-by-night than the actual trusted peer. Only exchangers will submit themselves to a normal accounting audit.

[casascius]

Steve's comment of putting in just a hash of all the unspent transactions got me thinking...

What if that was just one more thing that all miners verified?  If the txid's of all the unspent transactions were arranged into a (very large) merkle tree and the 32-byte root hash were saved - PER BLOCK - then this validation would be network-wide.

If that were done, clients could receive pruned blocks.  They would not necessarily know what was missing, but if they were able to calculate the same unspent merkle root hash with what remained, they would at least know that they had all the transactions anybody cared about.

[casascius]

The only participants of the bitcoin network that require trust are the exchangers. They will always have a cash flow to support the ancillary requirements of the network. The sooner you guys stop thinking about miners (and mining pools) as a centerpiece of the network, the quicker you'll arrive at a workable solution to this problem. Miners (of fresh bitcoins) need to be relegated to the status of peons. The actual value added is in the refining (collecting transaction fees) and the currency exchanging. Only exchangers have the incentive to maintain long-term business presence in the ecosystems. Pretty much everyone else is better served by being fly-by-night than the actual trusted peer. Only exchangers will submit themselves to a normal accounting audit.

Well, miners happen to also create the blocks that serve as the transaction record, so making sure that they can properly tell the difference between a valid transaction and an invalid one unassisted is relatively important.

[Steve]

Steve's comment of putting in just a hash of all the unspent transactions got me thinking...

What if that was just one more thing that all miners verified?  If the txid's of all the unspent transactions were arranged into a (very large) merkle tree and the 32-byte root hash were saved - PER BLOCK - then this validation would be network-wide.

If that were done, clients could receive pruned blocks.  They would not necessarily know what was missing, but if they were able to calculate the same unspent merkle root hash with what remained, they would at least know that they had all the transactions anybody cared about.

That's exactly what I was saying (though perhaps not so clearly).

[2112]

Well, miners happen to also create the blocks that serve as the transaction record, so making sure that they can properly tell the difference between a valid transaction and an invalid one unassisted is relatively important.

Again, long term miners just don't have place in the bitcoin ecosystem. Only refiners do the job you describe. Mining was sort of a temporary workaround for bootstraping the project. Satoshi was just very polite in programming an exponential decay of mining and leaving refining revenue to further modifications.

I used the word "peons" because in my experience it will force people to declare their intention and their psychological membership in the bitcoin caste system.

[bulanula]

This is a valid issue :

-adoption is difficult for people on dial-up or mobile broadband
-HDD gets very hot in laptops

[ctoon6]

This is a valid issue :

-adoption is difficult for people on dial-up or mobile broadband
-HDD gets very hot in laptops

dial-up, get real, it was hardly useful back in 2000 where almost all pages were bare.
most hard drives in laptops are now ssds.

we are moving forward in time, not back, so why would we directly cater to people who refuse to upgrade? i don't think 1 or 2 gigs is too much bandwidth to ask a mobile connection for, sure it will be slow, but thats a 1 time thing. but to keep it reasonable, you must keep the chain below a max of 2gigs, then id consider it "too big" to download for the average person.

[bitplane]

dial-up, get real, it was hardly useful back in 2000 where almost all pages were bare.
most hard drives in laptops are now ssds.

we are moving forward in time, not back, so why would we directly cater to people who refuse to upgrade? i don't think 1 or 2 gigs is too much bandwidth to ask a mobile connection for, sure it will be slow, but thats a 1 time thing. but to keep it reasonable, you must keep the chain below a max of 2gigs, then id consider it "too big" to download for the average person.

You're part of a tiny, privileged minority of people who live in well-connected countries where the cost of living is high and technology is relatively cheap. This isn't true of most of the world.

edit: http://www.netindex.com/value/

[casascius]

...but to keep it reasonable, you must keep the chain below a max of 2gigs, then id consider it "too big" to download for the average person.

That's the problem.  The chain grows without limit, and its growth rate increases exponentially.  It's not impossible that we could have a 2 gig chain this year.

[defxor]

I want to add that the amount of HDD "swapping" on new blocks is unreasonable. It sounds like the client performs defragmentation.

Can't this be optimized? Cached? It's very annoying and doesn't do any good for the disk.

I've seen constant write-to-disk of about 1MB (megabyte) per second on both old and fresh installs by multiple users. I thought naïvely this was a Mac specific 0.3.24 problem (edit: but 4.0rc2 seems to do it a lot less). Is this what you mean above?

Constant disk writing slows down a lot of systems, and I advice against running Bitcoin if you have either a laptop and/or an SSD because of it. I'd say it's actually more of a problem currently than the size of the blockchain (although the latter is almost a deal breaker when it comes to showing how Bitcoin works to new users: "Come back tomorrow, if you punch a hole in your firewall that is").