Resolution of simultaneous transitions

[tclose]

Currently in the specification it is left to the user to ensure that simultaneous transitions (i.e. within the same timestep) do not occur in the design of Dynamics classes. However, I have been thinking that this is not really practical in the design OnEvent transitions triggered by incoming events, which can happen at any time. Maybe we should require an explicit hierarchy in transitions to avoid such cases, as rare as they are likely to be.

[tangorn]

As far as I understand the issue the question is whether we process neuronal dynamics event before or after synaptic events if they happen simultaneously. This does not seem as an issue with alpha, beta, and even exponential synapses as they only affect voltage indirectly, but can be an issue with delta synapses. My first take is to process synaptic events first, neuronal dynamics events last within the time step. This way we will not have a case when a neuron that has just produced a spike has high voltage due to simultaneous input from delta synapse.