Yes The experimentation part is the most important one here.
So, there are very different possibilities:
Not set credit limits at all: each agent decides if they want to accept an incoming transaction based on the balance of the counterparty. I donāt know how much this pattern can be scaled, it seems to me that this works in trust networks when you roughly known one another.
Set the credit limits as properties in the DNA, they are fixed and canāt be changed, and apply to everybody. If you want to change those, you need to create a new DNA with new properties and migrate users if you need to. Updating this is hard because you need to maintain determinism of the validation rules, but maybe there is a way, with some governance mechanism.
Have a dynamic credit limit per each agent, according to a function that has as an input the source chain of the agent. This way you can say, every agent has a credit limit proportional to the amount they have contributing to hosting, in the case of holofuel. Or things like that.
Bear in mind that Iām not by any means knowledgeable in this area (hence me wanting to experiment ), other people could come up with much more interesting patterns.
Yeah so that would fall on category 2 but being able to dynamically set the limits, but we need some good validation rules to be able to handle thatā¦ Right now the only way I see involves comparing oneās timestamp with the timestamp of the last update of the global setting of the DHT, which means that there might be some accidents along the way when that setting has not yet made available in the DHT, and this could be attacked by setting your own timestamp before the decision was madeā¦ I dunno, we have to come up with smart patterns to go around these issues and maintain determinism of the validation rules.
Hey @guillemcordoba Enjoyed the discussion on this thread so far. With regards to your ādouble-spendā scenario, would it help if the credits transfer between counterparty or counterparties involved multiple steps, like what holo fuel is implementing?
This means, there will enough time (because there are a few steps in the process) for at least the 1st step to be recorded and for any node to check his counterpartyās chain for additional invalid recorded entries while not committing to go ahead (until all steps are done).
I think holo fuelās tx process involved more steps for higher security, but this may also increase ālatency/finality timeā. Am i understanding this correctly?
I think this problem occurs anytime you have asynchronous requests between the getting of the source chain of the counterparty. This is the case in my implementation, since I want for members of communities to accept/reject offers depending on the balance of the counterparty, so this needs to be asynchrounous (not in the same zome call). As far as I understand, holofuel does this synchronously, so they donāt have this problem.
I actually thought that all txs in holo fuel or holochain apps are done asynchronously. How is holo fuel able to do their steps synchronously ?
Do you mean steps can done in Holo Fuel synchronously in scenarios like (alice-bob & alice-carol) counter parties & public validations? cc: @pauldaoust@artbrock
What I meant is that HoloFuel sends the source chain and sign the transaction synchronously in the same call, not in two different calls. Actually I wouldnāt want to say too much about how HoloFuel works since I donāt understand fully the validation logic yet.
I followed your presentation with Grace at the Open2020 Coop webinar a week ago. I knew the Holochain since its inception but i am a sociologist (coordinator of the WiseLifeLab- University of Torino), so I wonder if we can have a short interview in which we can present our current proposal (timeline June 26th9 about Covid-19. We prepared a protocol based on tokenization but we would be glad to know if a non tokenized way is possible. My believe is that Holochain could work beyond the mutual credit game, but I am not sure about it.
Thank you for your reply.
Absolutely! Holo-REA will enable any kind of economic networking with any kinds of value flows, including mutual credit, but also food, seeds, biodigester inputs and outputs, energy, pie, whatever.
The experimentation part is the most important one here.
), other people could come up with much more interesting patterns.