On the main DAG? On a second layer? Or something else?
How does it work basically?
Add a comment
|
2 Answers
In IOTA timestamp accuracy cannot be trusted.
"The tangle is a graph with only a partial order structure, which makes it difficult (in fact, generally impossible) to establish the correct time order of transactions. Even if all transactions have timestamps on them, we cannot be sure that all these timestamps are accurate"
Source: http://iota.org/timestamps.pdf
In blockchain all blocks are organised into a linear sequence over time so this problem does not persist.
To make Smart Contracts on the Tangle possible, an "off-tangle" oracle will be implemented.
IOTA Foundation is actively working on Oracle platform for IOTA where the plan is also to enable true interoperability with all other Blockchain platforms.
-
So the main DAG will be left untouched, and a second layer of the Oracle platform will be implemented? Then Smart Contracts and Interoperability will be enabled on this second layer, am I understanding it correctly? Commented Dec 27, 2017 at 5:42
-
Ales has his own take, but I think he is perhaps ruling out smart contracts based on perceived functionality from other implementations.
http://iota.org/timestamps.pdf
Even if all transactions have timestamps on them, we cannot be sure that all these timestamps are accurate (there can be some malicious nodes that want to fool the network about the true time when their transactions appear, and/or some nodes with a wrong clock). Nevertheless, one can determine the confidence intervals for timestamps with reasonable accuracy
Timestamps per say are not important to implementing smart contracts. What is necessary, is a proper ordering of when transactions were completed. Something that datetimes are very good at, but it can be accomplished in other ways.
If that confidence interval approaches 100%, you can create logic that depends on that order.
Hell, we can't even be sure that a Transaction was even confirmed or not with a 100% confidence interval.
In order for IOTA to be successful, we need to be comfortable with the confidence intervals and probability distributions.
-
Judging from the link Ales provided, it seems clear to me that Smart Contracts will be implemented on a second layer - an Oracle platform, on top of the Tangle's base protocol. But are you proposing a solution to directly implement Smart Contracts on the Tangle? Commented Dec 27, 2017 at 5:46
-
Yes, precisely. The way the whitepaper about timestamps is written, it makes it sound like "yea, they don't work the way everybody is used to them working but it is possible". In a similar way that Michelin tires can "guarantee" that your tire will last for for $50,000 miles we can identify the confidence interval of a timestamp. 99% 99.9%, 99.99% or 99.9999% How many decimals is acceptable? Is it reasonable to achieve such confidence intervals? Those are the next questions to answer. I view the Oracle second layer approach as a fall back plan. Commented Dec 28, 2017 at 19:08