Do not forget that a transfer in a blockchain and a transfer in safe are, technically, completely different. in a blockchain any transaction must be retransmitted to all nodes while in safe only the sender group, the receiver, the one containing the MD and those that routing the message work.
Also in Safe the sender must sign and send each safecoin (or possible divisions). So a work must be done proportional to the amount of coins to send. This greatly hampers spam.
And if we came to the conclusion that spam is a real problem it could easily be avoided by asking for a little pow in every free transaction.
4 Likes
If the cost of a change of ownership is proportionate to / equal to the farming rate, the same mechanism should apply, with the farming rate being adjusted by the network to balance supply & demand of resources farmers provide (including managing ownership changes).
If demand spikes due to a very large volume of coin ownership changes, the network will sense excessive demand & the farming rate will be automatically adjusted to reduce demand and incentivise new supply.
Sorry if I’m missing something here - I’m not clear on many of the details!
2 Likes
This only occurs if a spammer or cracker tries to beat the system by doing bad things.
Like trying to multiple spend where they tell the network to transfer a coin to address 1 then immediately try to spend the same coin to other addresses before the network has a chance to consider the first. It is in these cases where the cracker/spammer is trying to transfer the coin multiple times before the first has a chance to complete.
So David mentioned that in this case the coin maybe lost to the spammer/cracker/cheat
Yea I realised after posting and after @fergish that I was a little bit turdish, and thanks for understanding. But yes the free + no spam was too much for my wee bit of brain power.
Data chains is after the first time David mentioned it, but If I recall correctly he has mentioned it in passing afterwards. I think its to do with how when a ownership transfer is requested that nodes that have a copy of the coin start changing the owner and while some have changed the owner another request is causing other nodes that have a copy of the coin to be changing it to another owner again. Now if its just twice then it’ll sort itself out. But if you do too many then in the final case the owner may not be agreed upon and thus the coin cannot be used again
3 Likes
Now this is where testing will be revealing a lot about the cost of a PUT. It may even be that a minimum safecoin amount will need to be set according to the relative size of the network. So while the network is small a higher minimum price is set and as the network grows the minimum will drop.
Also the price is dependent on the amount of spare resources. So as nodes leave then spare resources drop and the price will rise. As nodes join then the potential for spare resources to increase and a price drop.
So if the network cannot handle large amounts of operations then this means node count is lower and one expects the price is likely to be higher. As the network grows it will be able to handle more operations and in all likelihood the price will be lower. This will not be universally true but should be a good general rule of thumb.
I believe that the minimum price of a PUT as is suggested in the RFC is too small. The lowest price is on the order of 0.0000000000000000001 of a safecoin (1,000,000,000,000,000,000 PUTs per coin at minimum cost) And a PUT is upto 1MByte of data, believe me this is a very big number. (1,000,000,000,000 TBytes)
Even the 1000’s of billions is a little low of a price in my estimation since that could mean some one can store 1,000,000,000,000,000,000,000 bytes (1,000,000,000 TBytes) per safe coin if plenty of spare resources. In other words picking the right time to store data a person could just keep uploading over their lifetime for one safecoin. It needs discussions and I am reluctant to ask David to comment at this time since they are busy on other things and testsafecoin is in the future.
4 Likes
I voted Yes
At first cause of the spam attack.
Second I think it increase the value of safecoin if at least some amount of safecoin is needed in the network.
I don’t think it should be really expensive but I think it should not be too cheap aswell everyone should be able to pay 0.001 dollar per tx. On avarage.
Might be a little shallow but ethereum also gained a lot value cause of the tokens and the need of ethereum. Increased value brought a lot people to ethereum.
2 Likes
Exactly! We are talking of a pretty small data change here. The content is unchanged and it is just a switch of ownership. No data is moved, as the safecoin stays where it is.
Adding a transaction to a blockchain is a completely different kettle of fish. 1000s of nodes must confirm it and the number of confirmations per second is throttled by the block rate, which in turn is dictated by latency/propagation limits.
I hadn’t thought about the client having to do the signing too, but that definitely adds a resource cost to the process. While the cost to the network may be asymmetric and higher, it is also dispersed over the consensus group. If this results in each group node having <= the resource usage of the client doing the signing, I doubt spam will be an issue - the cost to the client vs the network will be prohibitive.
7 Likes
I shouldn’t be in this thread because I don’t have the slightest clue about most of its content but a quick skim of the language in the first 3rd makes it seem like SAFE is playing out a microcosm of some themes in economic history. (After release I’ll seek a primer to try to become educated in a basic way on its mechanism)
-
Fee sounds like a Tobin tax to check an externality.
-
Not having a fee seems to resonate with Islamic concept of no interest or even Masonic experiment interest free banking.
-
Somewhere was a reference to programatic auromous banking entity or banking free of manipulation and free of animal spirits- which irony Scotland has a deep experience with programed or free banking free of rate manipulations running at a set rate (something Milton Friedman once suggested) but the wealthy absolutely hated its apparently bad for concentrated wealth when they don’t have an entity they can call on to manipulate rates.
Thinking quantum computing may help with these transfer costs problems possibly on 4 fronts:
-
Q communication- tachyonic (ugh) presumably noiseless (ugh again) almost limitless bandwith connecting every node directly to every other or some close equivalent- the Q varient of decentralization makes it seem like everything is in virtual contact or super local. Could be sheer nonsense because I dont know what I am talking about and plenty of physcist say this stuff cant happen but current tech cutting edge seems to be moving this way with break through after rapid break through.
-
Q compute- to truely simulate the universe do you need a q system the same size or can you compress it into something the size of a refridgerator? Have a feeling it might be the latter because of a prospect in Q storage. There is serious speculation that an electron can either hold at least the library of Congress or an infinite amount of info. If its the infinite than its just digital physics and any Q computer is just a portal to the substrate. Regardless, having so much more in compute cycles without having a space or energy tax will have to do amazing things to the marginal transfer cost or transaction cost.
-
Q. Storage. If a single electron has infinite storage space you don’t have to solve one of the fundamental problems SAFE addresses plus it does great things to the transaction cost.
-
Q AI. AI even at the search level will be a great gate keeper and an incredible efficiency maximizer.
Again just the way a Q machine can solve a Hamiltonian to produce throughput, AI (Q aided) may get fine grained to the level of individual transactions.
Final silly point is nor surprised at all if SAFE is a software analogy for a q computer substrated for the universe i.e digital physics. But let us not forget its going abacus…calculator… classical computer… q AI learning machine… SAI and with each step it seems less machine like and more organic and the theory predicts we’re already surrounded by the nth step. To make this seem less like BS, pretty sure digital physics sees the universe using encryption and compression schemes in the behavior of matter.