Bitcoin Forum

Alternate cryptocurrencies => Altcoin Discussion => Topic started by: kuverty on June 09, 2012, 11:17:10 PM



Title: Making the work Bitcoin performs something useful?
Post by: kuverty on June 09, 2012, 11:17:10 PM
All right, I agree that Bitcoin is a wonderful and revolutionary idea. But the value it has, everyone understands it has some problems. Yeah, that hash-calculating solution is fine but Bitcoin could actually be valuable. It has it sides that the Bitcoin network does not work for any specific purpose than itself, but I think it could be used for something.

What is the work that miners do? They try to get a suitable SHA-hash from some random block so that it at the average requires a certain amount of work. But this work, this hashing, has no use to anybody.
It is as if I was to lift a rock ten thousand times and then get a certificate for doing it that I can exchange with others (peers could be watching me to confirm that I lift the rock 10000 times and get a little share for the confirmation effort).

Would it be too hard to set the goal of Bitcoin miners to be something different? I could imagine number-theoretic investigations, for example (even finding larger primes would be more useful). Proof of work would be harder, but doable.

I bet there's a lot of scientific/useful computing that could be turned into some kind of a goal like finding a block is. I could very well buy Bitcoins, if I knew the money was going to making progress towards a mathematical conjecture or modeling something suitable.

So, any comments? If people could just rent their computers like that, it would make the computing capacity of the world a lot greater - why waste expensive CPUs and GPUs and what else. Is there any project aimed towards this or reasonable discussion you can point me to?


Title: Re: Making the work Bitcoin performs something useful?
Post by: Revalin on June 09, 2012, 11:29:27 PM
https://bitcointalk.org/index.php?topic=52155.0;all
https://bitcointalk.org/index.php?topic=60703.0;all

tl;dr: You need to find a problem that is: 1) very hard to solve (about 20,000 GPU-hours per solution); 2) trivial to verify it was done correctly (less than 1 GPU-second per check); 3) politically neutral (we have to agree on the problem being solved, and some people will object to using their GPUs to solve some problems).

No one knows a useful problem in this category, so we're using SHA256.  Feel free to suggest something better, but criteria #2 is the hard one.


Title: Re: Making the work Bitcoin performs something useful?
Post by: kuverty on June 09, 2012, 11:31:11 PM
To add a point and make this thread look more interesting because it has a reply - I think this is the idea, if implemented (that anybody could rent their computer cycles for some value in a cryptocurrency like BTC, for some useful computations), would be the thing that would make cryptocurrencies a really big phenomenon. Big corporations and nations would buy cycles when needed and bring the price up. The Global Computing Network, and hooray for science and technology.

Ah, I see a reply there. Well, SHA-256 sure is well tested but there must be many more things that could be used if we have confidence in something, like Riemann Hypothesis... believing in SHA-256's invulnerability requires a bit of trust too. There must be a group of computer scientists somewhere thinking about this right now, but what should it be.

I'm not a hater, I'm a supporter. Not implying you meant that though.


Title: Re: Making the work Bitcoin performs something useful?
Post by: mcorlett on June 09, 2012, 11:34:41 PM
I think this is a loaded question. The work is not useless, because it helps to secure the transactions taking place on the network.

A new proof-of-work scheme needs to fulfill the following criteria:
  • The work must be easier to verify than to do.
  • The difficulty of finding a solution must be adjustable.
  • The scheme mustn't be prone to a "replay attack" of sorts, where you may use the same solution on a conflicting transaction or block.
Right now there exists no such thing, and I have a hard time imagining there will in the future.

Your example with prime numbers would fail because the difficulty is not flexible. Rather, it continually becomes more difficult.


Title: Re: Making the work Bitcoin performs something useful?
Post by: kuverty on June 09, 2012, 11:39:58 PM
So how about a little off-topic there, is there a nice way to rent some of your CPU cycles for someone? I think this fits in this topic. Something general-purpose, not just GIMPS or SETI


Title: Re: Making the work Bitcoin performs something useful?
Post by: Revalin on June 10, 2012, 12:15:14 AM
You probably mean like:  https://en.wikipedia.org/wiki/BOINC

But there's also:
https://aws.amazon.com/ec2/
http://www.bitvps.com/



Title: Re: Making the work Bitcoin performs something useful?
Post by: tatsuchan on June 10, 2012, 12:34:52 PM
So how about a little off-topic there, is there a nice way to rent some of your CPU cycles for someone? I think this fits in this topic. Something general-purpose, not just GIMPS or SETI

I JUST made a new thread about this in development topics.  I don't know how it works internally to a great extent, but I know that I mine namecoins along side my bitcoins.  How hard would it be to get a pool to start scouting out companies for extra work, and paying us more for our work?  What would developers have to do?

THIS idea could help the Bitcoin network as mining starts to pay total shit when halved enough, and people like me have small (2 ghash+) gpu farms, and no interest in buying $1000's of dollars worth of fpga's that have questionable potential to ever pay it back.

EDIT:/ Here is the topic I brought up earlier.  https://bitcointalk.org/index.php?topic=86684.0 
I want to see if there are ways to dedicate hash power to making money or/and donating to research.


Title: Re: Making the work Bitcoin performs something useful?
Post by: barbarousrelic on June 10, 2012, 12:46:42 PM
The reason that we can't use something like Folding@Home is that it would allow the operators of Folding@Home to potentially manipulate the blockchain. Individual nodes can't verify that a protein has been successfully folded on their own, they would have to trust the operators of Folding@Home. We would have to find some kind of useful calculation that everyone can verify without having to trust a third party.


Title: Re: Making the work Bitcoin performs something useful?
Post by: dree12 on June 11, 2012, 08:58:06 PM
The work Bitcoin does is useful, for breaking SHA-256 (after all, the miners are trying to find partial collisions). When it becomes financially viable, miners that develop better software or hardware and smarter semi-pre-image finders will earn more money. In the future, it may be possible to reverse some double-SHA-256 hashes that are near-zero. This work is just as useful as finding large primes.


Title: Re: Making the work Bitcoin performs something useful?
Post by: markm on June 12, 2012, 12:31:18 AM
This work is just as useful as finding large primes.

Primes don't change as fashions do though. SHA1 is kind of arbitrary, kind of old, and the closer our SHA1-reversal rainbow-table gets to complete the less fashionable SHA1 will likely become.

So no, nice try but a bit of a fail there. Primes are ridiculously fundamental, though a way to construct them instead of having to go searching for them would be real nice.

-MarkM-


Title: Re: Making the work Bitcoin performs something useful?
Post by: dree12 on June 12, 2012, 01:07:42 AM
This work is just as useful as finding large primes.

Primes don't change as fashions do though. SHA1 is kind of arbitrary, kind of old, and the closer our SHA1-reversal rainbow-table gets to complete the less fashionable SHA1 will likely become.

So no, nice try but a bit of a fail there. Primes are ridiculously fundamental, though a way to construct them instead of having to go searching for them would be real nice.

-MarkM-

  • SHA2 is significantly different from SHA1.
  • It is impractical to make a rainbow table; however, if we obtain a preimage for low hashed values, we can gain greater mathematical understanding of SHA2


Title: Re: Making the work Bitcoin performs something useful?
Post by: markm on June 12, 2012, 01:21:18 AM
Oh sorry did I type the wrong SHA-number? If so sorry, typo.

-MarkM-


Title: Re: Making the work Bitcoin performs something useful?
Post by: punningclan on June 12, 2012, 01:22:38 AM

It is as if I was to lift a rock ten thousand times and then get a certificate for doing it that I can exchange with others (peers could be watching me to confirm that I lift the rock 10000 times and get a little share for the confirmation effort).
Isn't that what Bricklayers do?  ???

Quote
I bet there's a lot of scientific/useful computing that could be turned into some kind of a goal like finding a block is. I could very well buy Bitcoins, if I knew the money was going to making progress towards a mathematical conjecture or modeling something suitable.
Many scientific problems exist that could benefit from giant super computers however none of them could ever be verified without some third party to check and this introduces a middleman whom everyone would have to trust, quite apart from the fact that they would show none of the regularity required by a currency.

Quote
I think this is a loaded question. The work is not useless, because it helps to secure the transactions taking place on the network.
+1
It secures the blockchain, heats my house and looks good on naked women what else do you want?

Like a one armed bandit all the mechanisms required for a random result aren't wasted they produce the result. Is the heat coming off my CPU wasted effort by my laptop?

It's really a trick of the light, in Bitcoin we actually already have the answer, the problem has already been solved and we know that it takes a certain amount of work to find a certain number of zeros. To do useful work you can't really know the answer. In a way the work that is being done is just testing whatever cryptographic engine Bitcoin happens to be using at the time.

This phantom problem will be solved as Bitcoin migrates to technologically better random number generators.
 


Title: Re: Making the work Bitcoin performs something useful?
Post by: Alex Coventry on June 12, 2012, 04:54:52 AM
I had an idea for doing this (https://bitcointalk.org/index.php?topic=64421.msg761836) a few months ago.  I've come up with a few improvements and hope to start working on an implementation soon.