ASIC

[rupy]

I think that if ASIC's are sold to BFL pricing/performance/delivery leads there will be a new coin (or bitcoin will change) that uses some other hashing, say SHA512 and the whole community will switch to that within a week?!

[Kluge]

Nope.

[Philj]

not going to happen. sha256 would have to be truly broken before bitcoin switches

[Meni Rosenfeld]

and the whole community will switch to that within a week?!

Um, no. Simple changes to the protocol take months. Difficult, compatibility-breaking, unjustified changes would take years. So would getting the entire economy to move to an alt coin.

ASIC equipment available to the public is good for Bitcoin.

[Garr255]

and the whole community will switch to that within a week?!

Um, no. Simple changes to the protocol take months. Difficult, compatibility-breaking, unjustified changes would take years. So would getting the entire economy to move to an alt coin.

ASIC equipment available to the public is good for Bitcoin.

Only if it is truly available to the public, or if whoever receives it first chooses to use it for the good of Bitcoin.

[rupy]

and the whole community will switch to that within a week?!

Um, no. Simple changes to the protocol take months. Difficult, compatibility-breaking, unjustified changes would take years. So would getting the entire economy to move to an alt coin.

ASIC equipment available to the public is good for Bitcoin.

We'll see, atleast I know I'm going to use my FPGA's for something so it will probably be some kind of hashing... I will build my own coin if thats what it takes...

You are so sure that it's good to lock us into SHA256?

[Dargo]

and the whole community will switch to that within a week?!

Um, no. Simple changes to the protocol take months. Difficult, compatibility-breaking, unjustified changes would take years. So would getting the entire economy to move to an alt coin.

ASIC equipment available to the public is good for Bitcoin.

We'll see, atleast I know I'm going to use my FPGA's for something so it will probably be some kind of hashing... I will build my own coin if thats what it takes...

You are so sure that it's good to lock us into SHA256?

The SHA-3 winner will be announced later this year. It would be cool to build a coin on that. I think it's going to be quite a while before SHA-256 is compromised, so Bitcoin is fine for now. But probably a switch to something else will be needed eventually.

http://www.drdobbs.com/security/finding-the-new-encryption-standard-sha-/231700137 

[Meni Rosenfeld]

and the whole community will switch to that within a week?!

Um, no. Simple changes to the protocol take months. Difficult, compatibility-breaking, unjustified changes would take years. So would getting the entire economy to move to an alt coin.

ASIC equipment available to the public is good for Bitcoin.

We'll see, atleast I know I'm going to use my FPGA's for something so it will probably be some kind of hashing... I will build my own coin if thats what it takes...

You are so sure that it's good to lock us into SHA256?

ASIC doesn't lock Bitcoin into SHA256, it just means that switching to a new hash function should be gradual over several years.

[rupy]

Eh, ASIC totally locks the whole global bitcoin mining equipment into SHA256, which will be millions of dollars worth of hardware. Say somebody breaks SHA256 and 99% are mining with ASIC then bitcoin is dead.

Switching gradually? How do you suggest doing that?

EDIT: We should move towards making bitcoin MORE flexible/scalable in the core (change hashing algorithm (on the fly?) and transaction speed) instead of focusing on stuff like BIP16/17...

[Transisto]

Eh, ASIC totally locks the whole global bitcoin mining equipment into SHA256, which will be millions of dollars worth of hardware. Say somebody breaks SHA256 and 99% are mining with ASIC then bitcoin is dead.

Switching gradually? How do you suggest doing that?

EDIT: We should move towards making bitcoin MORE flexible/scalable in the core (change hashing algorithm (on the fly?) and transaction speed) instead of focusing on stuff like BIP16/17...

If there is a significant risk of someone breaking SHA256 ... then ... you could start a new thread about it, hopefully with a descriptive title.

[rupy]

You're missing the point, ASIC's are good for bitcoin if bitcoin was perfect. But it needs to change and silicone doesn't change. You need to keep it in software until it can scale and is prooven. Going ASIC now is a poor choice. We're only at the beginning of crypto currencies.

[rjk]

You're missing the point, ASIC's are good for bitcoin if bitcoin was perfect. But it needs to change and silicone doesn't change. You need to keep it in software until it can scale and is prooven. Going ASIC now is a poor choice. We're only at the beginning of crypto currencies.

Please describe the imperfections in Bitcoin that make this a bad idea.

[rupy]

1) You need to be able to switch hashing algorithm and block chain for proof of work only on the fly. Allowing for ANY and ALL kinds of crypto currency on the same platform.
   b) You should build in "p2pool" into the protocol to avoid pooling.

2) Transaction scalability sucks, basically the block chain shouldn't record transactions, but the account balance should be stored in the p2p cluster but not in the block chain. See https://bitcointalk.org/index.php?topic=91397.0.

These two above should be 2 completely separate applications. One for mining and one for transactions. (of course the mining application will need the transaction application to work but not the other way around, which is the whole point!)

Using a block chain as proof of work with an adaptive difficulty is a good and pretty proven concept though proving the rarity needed for any store of value.

[Xenland]

  b) You should build in "p2pool" into the protocol to avoid pooling.

Just something odd about this type of assumption, but I don't know enough about ASIC's to really give an adequate answer.

I'm going to have to respond with, What nobody agrees with p2pool integrated into the protocol and still use the current one? 

[rupy]

  b) You should build in "p2pool" into the protocol to avoid pooling.

Just something odd about this type of assumption, but I don't know enough about ASIC's to really give an adequate answer.

I'm going to have to respond with, What nobody agrees with p2pool integrated into the protocol and still use the current one? 

This has nothing to do with ASIC, but just a side flaw of the current bitcoin implementation.

[Bitcoin Oz]

I would support a coin based on the new SHA-3 winner. While you are at it implement transaction pruning ftw.

I mean if bitcoin is going to move to it eventually why not prepare the road ?

[Meni Rosenfeld]

Eh, ASIC totally locks the whole global bitcoin mining equipment into SHA256, which will be millions of dollars worth of hardware. Say somebody breaks SHA256 and 99% are mining with ASIC then bitcoin is dead.

Hash functions aren't broken overnight. When cracks start to be found in SHA256 we can worry about this.

Switching gradually? How do you suggest doing that?

Change the protocol so a given % of blocks needs to be SHA256 and a different % some other function, and change the percentages gradually. This gives hardware owners less to worry about, and ample time to develop ASIC for the new function.

EDIT: We should move towards making bitcoin MORE flexible/scalable in the core (change hashing algorithm (on the fly?) and transaction speed) instead of focusing on stuff like BIP16/17...

We should focus on stuff that are relevant now, and defer stuff that will be relevant in decades.

1) You need to be able to switch hashing algorithm and block chain for proof of work only on the fly. Allowing for ANY and ALL kinds of crypto currency on the same platform.

Compatibility-breaking changes aren't made on the fly.

  b) You should build in "p2pool" into the protocol to avoid pooling.

But you have p2pool, so why is it so important to change the protocol?

2) Transaction scalability sucks, basically the block chain shouldn't record transactions, but the account balance should be stored in the p2p cluster but not in the block chain. See https://bitcointalk.org/index.php?topic=91397.0.

These two above should be 2 completely separate applications. One for mining and one for transactions. (of course the mining application will need the transaction application to work but not the other way around, which is the whole point!)

Using a block chain as proof of work with an adaptive difficulty is a good and pretty proven concept though proving the rarity needed for any store of value.

I don't think you understand how Bitcoin works. Proof-of-work is what secures transactions from double-spending.

[Xenland]

This has nothing to do with ASIC, but just a side flaw of the current bitcoin implementation.

Troll alert...... This has nothing to do with ASIC just Bitcoin? (Yet the thread is titled ASIC)


Do you understand the definition of an open-protocol?