Poll: Should SAFE Network tokens have txcosts?

Maybe, but the network has no time metric to measure against, and introducing such would be all sorts of headaches, and monitoring to enforce would introduce all sorts of security issues.

I’m for all mutations, including safecoin transfer, being one PUT.


The elephant in the corner is WHY.

Why oh why do we need free token transactions? The network has to do work for them and you want to make that free?

The costs for a owner change of a token is going to be Farming-Rate and that for the most part will be less than 0.00000000001 of a safecoin. In other words you will likely be able to do 1000’s of billions of owner changes per safecoin when the network has enough spare resources (farming rate very low)

SAFEcoin of course from a attraction point of view will be free, but since it costs safecoin and there will be network lags limiting the number of transactions a spammer can do.But even then it might cost a PUT whenever you send safecoin (no matter how many) for the receiver since they will be updating some wallet database. So a spammer will incur costs to spam SAFEcoin.

But now you want tokens completely free? You know that is the easiest thing to spam. Just pay the 1/1,000,000,000,000 of a safecoin to send the token and prevent the massive spamming

I am so sorry but I cannot stop laughing at this thought. Of course it will be spammed and spammed to death. Thats why we have the invite system, because the network was spammed on just a test net. What do you think will happen if you provide an attack vector the size of the pacific ocean for the spammers. Of course they will spam and spam hard, even if just to make a point.

As soon as you come up with a method it will then remove the attraction of “free”. As soon as you make it harder to spam you remove the attraction of free because now they have to do some sort of work. No need to consider the technical side, it is just obvious. Doing something is a cost of some sort.

to me the miniscule cost for transfers on the SAFE network will be the attraction and the difference with free for valid/real users of the token is so small that its hardly measurable. Just running a vault and geting one safecoin will likely pay a lifetime of fees for token transfers for the user. Exception is if the person only transfers when spare space is very low.

SAFEcoin has other measures to prevent spamming. For one you have to store your coin addresses somewhere and that will cost PUTs to do the list update. So spammers will have to be paying a number of PUTs every time they receive a safecoin so spammers will eventually use up their coins. So it will cost them dearly to spam the network by send safecoin.

But if you make tokens transfers completely free then this attack vector is not like safecoin but a completely free window for the spammers to attack the network with as many machines as they can.


Actually for people to receive safecoin there will have to be some PUT costs because they have to update their list of coins in their wallet and this will cost. For proper use of SAFEcoin this will be such a small cost but for a spammer they will eventually run out of SAFEcoins.

Oh and a balance method is the most likely solution to divisibility where the wallet fraction balance holds a maximum of just under one safecoin.


@neo has basically said all, and in a very good way. But I have some other angle to chime in with too :wink:

I think you are doing the same thing here, saying “they do it for a reason”. We should skip that all together, needs to get to the basics of it.

What is the problem, what needs to be solved?

I think the problem is not to verify if people would love to have something they want, for free. You would have to look for a quite contrived edge case to come up with an example when this would not be so.

The problem is something else.

That is just not correct. The only way would not be free transactions. Anything that is substantially cheaper, will compete.
So, it does not have to be absolutely, 100% free.
It’s enough that it is substantially cheaper.
(If we are comparing those things only, disregarding all other possible incentives.)

But there still needs to be some friction.

And actually, you come to the conclusion yourself:

And how does the network treat information? PUT costs.

Coming up with side cases for different types of information is clunky and brings about all sorts of headaches.

No don’t… treat it elegantly, and do not introduce side cases for different type of information.
Information is treated like any information, and we have less complexity and more robustness which equals longevity.

And again, tthis:


Well each coin will have a limit already on how many transactions can occur. Just the lag time involved in the consensus mechanism will be a limiting factor and is likely to be 1 second or so anyhow.

David also mentioned that if you try to operate (too many times) on the coin before this have completed it is possible the coin will be lost to the user transferring it because of mechanism.

As @fergish says there is no concept of time for the network.

@fergish if we charge for safecoin transfers then we get a chicken and egg problem. How does one pay for the first purchase of resource to pay for the payment to buy resource.


I was initially with @fergish here, that even safecoin should cost, however small the cost would be.

But, safecoin is a proper thing in the network, it has a function that no other thing will have. It is the fuel.
So, that is what earns it it’s special treatment. It is part of the network implementation, a fundamental part.

Tokens are just some information that users of the network decides to work with. It’s just a use case among others.

(Now, safecoin can probably be designed to work with fees too, it would just then have some other characteristics)

Not that I would prefer it, but it seems we already have this situation, (and there was some idea of inviting users by donating a small sum enough to pay for registration).


This is a barrier that could be as good as anyone. (if it holds true with a large enough probability)

Someone will push and push, and make careful measurements of preconditions and timings, and find out just how much they can operate on it to stay within x % probability of not losing the coin.

And then they will stay within that limit. And so will everyone else who learns of it.


That is a different one again. I was referring to the case where one has a safecoin and how does one pay for resource if you require resource to pay for the payment (transfer). Obviously you need a special case of owner transfer for paying for resource when one has no resourse. But again this is creating more special case code rather than reusing the safecoin transfer code and just have the destroy coin at the end.

The one you refer to is how does one get a safecoin in the first place. One answer is set up farming to earn a coin first, but this also requires free transfers otherwise special case code again for no coins and just earnt a coin.


Okay @neo and @oetyng, I concede the point re safecoin, though I’m open to that option. Pay to PUT for all storage, mutation or ownership transfer, except safecoin (probably).

@19eddyjohn75 – I know @neo was a bit chiding, but doesn’t this make sense?


Specially on SAFE,
a token should have at least the minimum fraction of its unit as fee.

If a token is divisible to 8 places after the decimal point the fee
should be a minimum of 0.00000001. Think 1 sat, for example.

As pointed out, fees prevent spam & manipulation & endless loops.
At the minimum, it serves also as counter for the total amount of tx.

1 Like

fees give incentive and value to anything…

Specially on SAFE the fee is in SAFEcoin for the PUT cost. And this mitigates against spam.

Any other fee in terms of the token is up to the token usage and procedures, but is not necessary for SAFE.


Sure, but there may be things to measure against - it doesn’t need to be a fixed time in seconds or minutes. There just needs to be something to throttle spam to neutralise its impact.

The cost to the network should be relatively low for moving safecoin about, so I am not concerned with network cost per operation (general PUT cost can subsidise it). What we don’t want is people killing the network through spamming something free of charge.

Wow honestly I didn’t know that it was that dirtcheap, that is freeish. Sorry for my ignorance.

Your right the free transaction is nonsense, but 0.00000000001 of a safecoin would position the SAFE Network tokens to be preferred in the future. This is way better than 0.0001 per tx on bitcoin and this cheap even enables us to compete with banks with SAFE Network tokens.

Huh tokens can be lost? Mindblown, please explain and is there a possible solution to this?

@neo is just straight to the point, what I say is idiotic and makes no sense (free + there will be no spam: LOL), he’s right and he luckily gave me more insight in the matter. The counters to my, make it free makes perfect sense.

Thanks again for the feedback @neo @oetyng @DavidMc0
@fergish I would pay for a more in depth talk about SAFEcoin SAFE tokens again, but this time a little more technical if possible.



I think the SAFE protocol isn’t meant to solve the problem of sending FREE data. It’s meant to build a autonomous web infrastructure which gives us some basic guarantees on our data because it’s incentivised. If you make any type of data transfer completely free, it WILL be exploited.

I think we can solve this problem by using a different protocol such as dat, which is already implemented as default in Beaker Browser, on top of which the SAFE Browser is built on, in conjunction with MD. And if the SAFE devs don’t plan on enabling dat, than we can still use dat over WebRTC which definitely will be available.

Dat is not incentivised, it’s a free network like Torrent, thus enabling FREE usafe. That has advantages and disadvantages of course, there are always tradeoffs.


Good idea. Topic is now in the queue.


About this isn’t datachain suppose to prevent data lost?

I’ll pay a 100 euros for an in depth version of this topic (sorry I feel that it should be at least a 1000 euros, but I’m balling at a budget at the moment). It might also get the wider crypto community looking better at what SAFE tokens can do for their project, thanks again @fergish :stuck_out_tongue:


Raise the price! :wink: I am not comfortable with 1000’s of billions of owner changes per safecoin, given the number of PC’s, scattered all over the world, that are involved in validating that ownership change. It still seems like a pretty low cost attack vector.

How about 10’s or 100’s of thousands of owner changes per safecoin?

Let’s pretend bitcoin is still $6000 and let’s pretend you can send it for $0.50. That means after you send that bitcoin 12,000 times, it’s gone. Can we stay within an order of magnitude of this? How about one order of magnitude bigger, so on the order of 120,000 ownership changes per safecoin? It’s still cheap, but not “invite spam attack” cheap (hopefully)

The network will set the cost per put (and therefore Safecoin transactions cost) dynamically based on demand and supply of resources.

If working correctly, the network will respond to an attack that increases demand to a level that could exhaust available resources by raising the put price until demand is no longer outpacing supply.


I think he was referring to the network load potentially incurred by having OWNERSHIP change be so cheap. IIRC only storage flood attacks are mitigated by what you posted above.