The key point is that all ATs would need to be *restored* to the previous state they had at the point of the re-org (and then re-executed).
Thanks,
Just to be clear, this is network based and not oracle based right (not an n of m signers trust model?)...
If it is completely trustless then how can you deterministically execute scripts within known times as per spec:
"It can be expected that over time an increasing number of ATs will be created so the cost for running AT code
must be kept to an absolute minimum. This will require AT code to be deterministic, and that its instructions
must not be able to take arbitrary amounts of time. This will limit the set of API functions accessible to an
AT to only those that do not require arbitrary scans over the blockchain (i.e. they should be only applicable
to indexed information)."
If its network based can you have cases where AT's are running trying to access info that is not indexed on the local node it is running on?
Also just to be clear, it is allocating/checking any memory prior to execution right? So its a load/store machine rather than a two address machine?