I love this kind of german accent... I would like it if you could drop a link to the web page so I can soak up the essential ideas that this 'tauchain' is based on.
The whitepaper is a bit outdated. Ohad has switched the logic to Monadic Second Order Logic (MSOL). In order for Tauchain to achieve it's goal, it must have in-built decidability. Ethereum does not have this which is part of the reason for why the DAO couldn't be secured before being tested by users.
oh, ok... he sounds like a few germans I knew from australia, well, a lot anyway.
interesting concept... I had similar ideas about a semantic network system years ago, when I also started working on my Spacetime Expansion Hypothesis. This is a very interesting progression on from the pure machine-language distributed computation of ethereum. I get the feeling that this is not so much a big money thing for a while yet, this will take a lot of development. More or less, the way I understand it, it's like a meta-distributedprocessing-blockchain system. Probably it would be too process intensive just yet but on the other hand... advanced AI processing could become a PoW system down the track, not too far.
oh you keep updating it :) good! I'm glad I commented now. This is very interesting stuff, indeed, although I think it's still a long way from where it will directly impact in the marketplace. More or less a system for teaching machines to understand human communications networks and assist them, right?
I think it has potential to be much bigger than Ethereum but Ethereum has good marketing and is more of a "cool new thing" even if it's not technically impressive to some of the people who studied it. I've said prior to TheDAO that Ethereum will not be secure, and so did others, and the main reason is it's Turing complete, it's not decidable, it's built on a house of cards.
You can't trust billions of dollars to Solidity. When you want to do an Internet of Things you actually need verified code, correct by construction, and if things are decidable you can have correct code which you can then limit the unexpected behaviors of certain processes. This adds a kind of concreteness which you can't ever get with Turing complete Ethereum.
Tau is a programming language. Tauchain is the platform. Agoras is the intelligent market. Together they will be great, and much more powerful than Ethereum. Ethereum might be more popular, because hackers prefer the technology (gee I wonder why?), but it's not what I would choose for projects which require serious security like voting or handling billions of dollars, or even the IoT. For mission critical code, you need correct by construction and I think Synereo is going to be more advanced than Ethereum and because of design decisions I do not think Ethereum can evolve into something new while remaining Ethereum, they need to ditch Turing completeness and Solidity in my opinion.
I did a lot of study of computer science and languages, and computer language processing back in the late 80s when I was a pre-teen. The field has a long way to go and it means it will take time.
I look at all the current fads in computer science, and basically, nobody is really doing the proper analytical stuff that is needed. Languages are blossoming like mushroom flushes but the important thing: Correctness - is not. A lot of computer languages let you quickly express a logic but they don't make it very clear what the logic does. This is where most bugs come from.
Like you say, without correctness, a computer system is dangerous.
Having said that, Steem is taking a very interesting approach by starting with simple stuff and more or less, and they don't say it straight out, but the blog is actually using human brains to mine. I actually came up with a phrase for this: Consensus Mining. I think that while machine processing is very important, in the immediate term, perfecting ways to build human consensus is going to be critical in governance. Dan has said and I have been also saying for some time, that with the right network infrastructure, humans can govern their economic interactions with each other. This is in my opinion the most critical immediate step, without this, projects like Tauchain won't even get noticed.
Typechecking can help for a start but hopefully Tau can bring better and more secure programming tools to the decentralization community. What good is decentralization if it doesn't improve security?
Tauchain is very hard to build the compiler part which has to be built correct by construction with no room for error or bugs. Once the hard part is built, then you because it's decidable it makes building apps much cheaper than on Ethereum. Ethereum development gave people the illusion of being easy and cheap but for any serious development it will cost you millions of dollars and still might be insecure. Toy apps on Ethereum can be developed for cheap and might even get some people to buy tokens but it wont be secure and so the costs do exist but are hidden in the cost of auditing the security and testing.
It is actually easier to develop secure code in C++ than to work with Solidity right now. Try it for yourself if you don't believe me. Ethereum at this time is mostly vaporware and toy apps.
I totally understand what you are saying. I come from a background in BASIC and assembler. Correctness is difficult to express in algorithms, it's definitely a branch of AI and a very neglected part of CS - because vulnerabilities tend to come from bad logic. A simple example that I pointed out in the github about the steemit.com website code - at first the tag field was verifying tokens just by counting spaces!
This is awesome, is there any role for Steem/Steemit to play in this? For example, a forum for proposing and discussing tauchain projects and monetarily rewarding developers?
I love this kind of german accent... I would like it if you could drop a link to the web page so I can soak up the essential ideas that this 'tauchain' is based on.
He is not German, he is Israeli.
https://github.com/naturalog/tauchain
http://www.tauchain.org
http://tauchain.org/tauchain.pdf
The whitepaper is a bit outdated. Ohad has switched the logic to Monadic Second Order Logic (MSOL). In order for Tauchain to achieve it's goal, it must have in-built decidability. Ethereum does not have this which is part of the reason for why the DAO couldn't be secured before being tested by users.
oh, ok... he sounds like a few germans I knew from australia, well, a lot anyway.
interesting concept... I had similar ideas about a semantic network system years ago, when I also started working on my Spacetime Expansion Hypothesis. This is a very interesting progression on from the pure machine-language distributed computation of ethereum. I get the feeling that this is not so much a big money thing for a while yet, this will take a lot of development. More or less, the way I understand it, it's like a meta-distributedprocessing-blockchain system. Probably it would be too process intensive just yet but on the other hand... advanced AI processing could become a PoW system down the track, not too far.
oh you keep updating it :) good! I'm glad I commented now. This is very interesting stuff, indeed, although I think it's still a long way from where it will directly impact in the marketplace. More or less a system for teaching machines to understand human communications networks and assist them, right?
I think it has potential to be much bigger than Ethereum but Ethereum has good marketing and is more of a "cool new thing" even if it's not technically impressive to some of the people who studied it. I've said prior to TheDAO that Ethereum will not be secure, and so did others, and the main reason is it's Turing complete, it's not decidable, it's built on a house of cards.
You can't trust billions of dollars to Solidity. When you want to do an Internet of Things you actually need verified code, correct by construction, and if things are decidable you can have correct code which you can then limit the unexpected behaviors of certain processes. This adds a kind of concreteness which you can't ever get with Turing complete Ethereum.
Tau is a programming language. Tauchain is the platform. Agoras is the intelligent market. Together they will be great, and much more powerful than Ethereum. Ethereum might be more popular, because hackers prefer the technology (gee I wonder why?), but it's not what I would choose for projects which require serious security like voting or handling billions of dollars, or even the IoT. For mission critical code, you need correct by construction and I think Synereo is going to be more advanced than Ethereum and because of design decisions I do not think Ethereum can evolve into something new while remaining Ethereum, they need to ditch Turing completeness and Solidity in my opinion.
I did a lot of study of computer science and languages, and computer language processing back in the late 80s when I was a pre-teen. The field has a long way to go and it means it will take time.
I look at all the current fads in computer science, and basically, nobody is really doing the proper analytical stuff that is needed. Languages are blossoming like mushroom flushes but the important thing: Correctness - is not. A lot of computer languages let you quickly express a logic but they don't make it very clear what the logic does. This is where most bugs come from.
Like you say, without correctness, a computer system is dangerous.
Having said that, Steem is taking a very interesting approach by starting with simple stuff and more or less, and they don't say it straight out, but the blog is actually using human brains to mine. I actually came up with a phrase for this: Consensus Mining. I think that while machine processing is very important, in the immediate term, perfecting ways to build human consensus is going to be critical in governance. Dan has said and I have been also saying for some time, that with the right network infrastructure, humans can govern their economic interactions with each other. This is in my opinion the most critical immediate step, without this, projects like Tauchain won't even get noticed.
Typechecking can help for a start but hopefully Tau can bring better and more secure programming tools to the decentralization community. What good is decentralization if it doesn't improve security?
Tauchain is very hard to build the compiler part which has to be built correct by construction with no room for error or bugs. Once the hard part is built, then you because it's decidable it makes building apps much cheaper than on Ethereum. Ethereum development gave people the illusion of being easy and cheap but for any serious development it will cost you millions of dollars and still might be insecure. Toy apps on Ethereum can be developed for cheap and might even get some people to buy tokens but it wont be secure and so the costs do exist but are hidden in the cost of auditing the security and testing.
It is actually easier to develop secure code in C++ than to work with Solidity right now. Try it for yourself if you don't believe me. Ethereum at this time is mostly vaporware and toy apps.
I totally understand what you are saying. I come from a background in BASIC and assembler. Correctness is difficult to express in algorithms, it's definitely a branch of AI and a very neglected part of CS - because vulnerabilities tend to come from bad logic. A simple example that I pointed out in the github about the steemit.com website code - at first the tag field was verifying tokens just by counting spaces!
Well my mind is blown
This is awesome, is there any role for Steem/Steemit to play in this? For example, a forum for proposing and discussing tauchain projects and monetarily rewarding developers?
This post has been linked to from another place on Steem.
Learn more about linkback bot v0.4. Upvote if you want the bot to continue posting linkbacks for your posts. Flag if otherwise.
Built by @ontofractal