Hi all,

I'm the author of that "Tangle" paper. Answering:

1. To approve means to reference or to confirm; that is, you have to cite these 2 tx' in order to issue yours. It is a good strategy to check if the 2 tx's you reference are not in conflict; otherwise, you risk that others won't approve your tx.

2. similar to bitcoin, i.e., finding a nonce that gives hash with N zeros in front

3. correct.

4. yes, exactly

5. yes. We are modifying the approval strategy, though, as the current design can be gamed.

6. yes (modulo the previous comment)

7. yes. Against (b), there is no other defence but to cap the tx own weight (or even set it to constant)

--------------

Does the DAG get pruned?

Do transactions have inputs and outputs similar to Bitcoin?

CfB is a better person to answer that.

in the event that two transactions reference the same inputs, the one with the larger cumulative weight wins?

yes, exactly