Do you have a background in distributed systems? In the same why the common advice with cryptography is, "one can build cipher unbreakable to the creator but not the world", the same way too "one can create a 'blockchain' that's unbreakable to the creator but not the world".
Git cannot provide byzantine fault tolerance. The fault tolerance it can provide is costly in human time.
Senior Engineer: Hi, I'm new here and have built a better blockchain. Let's use a chain of hash digest. It's way more efficient than PoW. It only costs .001/tx
CTO: But we need decentralized time stamping not the smaller set of always-online tamper detection.
SE: It's ok we'll be able to tell if the state updates change by monitoring the chain of hash digests. :>
[Months Later]
SE: Well, some attackers knocked our machines offline.
CTO: How do you know that the state updates between when the attack happened and when you got back online are valid and uncontested? You built some stupid slashing mechanism and if we publish invalid state our business goes under.
SE: We'll I'll have to go in to audit the logs, hopefully find some other people who have logs, and write software to see if maybe we can detect anything.
CTO: Isn't that going to blow up your transaction costs?
SE: I estimate 30k in total costs to do the audit.
CTO: so over our 800k customers that'll increase your tx costs to .038/tx, your hash-digest solution doesn't seem to be a cheaper alternative.
SE: I'm quitting to launch an ICO
CTO: Thank god.
Basically all the ICOs --Kik included-- are fraud: they won't be able to achieve what they claim to be able to do (governed by unavoidable physics, CS, cryptography, DS constraints) and punters have a myopic belief of what they can do. They gain market share by publicizing Bitcoin's very specific limitation. Bitcoin continues to chug along, is very clear about what it can and cannot do.
It's painful to have an appreciation for DS and a view into the cryptocurrency space, in the same way it's painful for a IETF network engineer to have the internet ossify around TCP and UDP. Yes there are drastically better CS primitives to engineer these systems with but the switching costs (social and economic) of transitioning a network are higher than the benefits.
You don't need byzantine fault tolerance within an organization. You need at most regular signing. For an example of how this works in the real world, consider the French NF525 system where immutable transaction auditing is legally required. Unsurprisingly, it's not done with blockchain proof-of-waste.
>You don't need byzantine fault tolerance within an organization.
Maybe you haven't encountered a need for BFT in organizations that you've observed but it would be fallacious to say "...and ergo no organization would need BFT" (round-trip fallacy). I agree that an org doesn't need BFT if:
-Their hardware never fails
-Their hardware never goes offline
-Their permission structure never changes / No (permissioned hardware or employee) ever become unpermissioned at the human layer.
-Their hardware is in critical time sync with other system all the time.
-...
I have yet to see such a system but do see many people who don't realize their DS needs. That's fine, most people don't understand the delayed costs in not properly auditing unstable logs. Someone's got to work for bitcoin in the future.
Git cannot provide byzantine fault tolerance. The fault tolerance it can provide is costly in human time.
Senior Engineer: Hi, I'm new here and have built a better blockchain. Let's use a chain of hash digest. It's way more efficient than PoW. It only costs .001/tx
CTO: But we need decentralized time stamping not the smaller set of always-online tamper detection.
SE: It's ok we'll be able to tell if the state updates change by monitoring the chain of hash digests. :>
[Months Later]
SE: Well, some attackers knocked our machines offline.
CTO: How do you know that the state updates between when the attack happened and when you got back online are valid and uncontested? You built some stupid slashing mechanism and if we publish invalid state our business goes under.
SE: We'll I'll have to go in to audit the logs, hopefully find some other people who have logs, and write software to see if maybe we can detect anything.
CTO: Isn't that going to blow up your transaction costs?
SE: I estimate 30k in total costs to do the audit.
CTO: so over our 800k customers that'll increase your tx costs to .038/tx, your hash-digest solution doesn't seem to be a cheaper alternative.
SE: I'm quitting to launch an ICO
CTO: Thank god.
Basically all the ICOs --Kik included-- are fraud: they won't be able to achieve what they claim to be able to do (governed by unavoidable physics, CS, cryptography, DS constraints) and punters have a myopic belief of what they can do. They gain market share by publicizing Bitcoin's very specific limitation. Bitcoin continues to chug along, is very clear about what it can and cannot do.
It's painful to have an appreciation for DS and a view into the cryptocurrency space, in the same way it's painful for a IETF network engineer to have the internet ossify around TCP and UDP. Yes there are drastically better CS primitives to engineer these systems with but the switching costs (social and economic) of transitioning a network are higher than the benefits.