I think this requirement is always met because once you can exchange turtles for Safecoin you have a route to fiat.
Err we canât do that. Because people would game the system by arbitrage trading Safecoin <-> (SAFE GB) to either gain more GB or more Safecoin, which will create a mess for the Networks supply/demand balance. That is why it has to be one-way conversion only.
Safecoin >>> SAFE GB
That is a very valid point.
People still have a hard time understanding Bitcoinâs fiat conversion value as well as the other cryptos. And I acknowledge it in my original statement. Right now, most people donât know how to value resources but maybe in the future everyone will understand (SAFE GB) as a universal term.
How much is that $10 shirt?
Oh itâs 166 (SAFE GB)
Two things I donât understand thenâŚ
- how does turtles->Safecoin cause a problem. Arbitrage itself is just how exchanges keep in step, can you explain what you mean by gaming through arbitrage. Why this serious enough to ruled out turtle->Safecoin exchange?
- if you can exchange turtles for fiat, how is that different from turtles->fiat->Safecoin and why is this ok when turtles->Safecoin is not?
Agreed. Perhaps serveral alternative methods for âSafecoinâ can be presented in the testnet and a play-market setup to see how people go buying, selling, making micro and mega payments etc under various regimes for various play-services.
Ask a non-technical use if 166 SAFE GB worth more than 166 SAFE TB and they will have no clue or be easily confused/duped. Almost all will understand that $Safe 16.6 is worth 10 times less than $Safe 166.0 though. Usability and familiarity are hard things to tease apart.
1 Like
You may be right, and it wonât matter.
We should test it, to see how people will react if we go down this road.
The arbitrage trading was for fiat profit motivation. But in terms of destabilizing the Networks supply/demand incentive⌠I donât know for sure.
I think a small Network would be vulnerable to storage capacity manipulation. The temptation for large farmers to collude in turning off their vaults is too tempting. Theyâd drive up the SAFE storage price, switch GB to SC to make sweet fiat profit⌠worst case scenario.
My understanding of arbitrage is that it is a normal and necessary market function, which is motivated by the profit it allows, which in turn is limited by competition between arbitrage traders.
Sounds good then.
Turtles are go! 
Yes, I think we should call them turtles, (Iâd suggest Yamanakaâs but its too hard to say). Or maybe Davids? There you and @dirvine share the acknowledgement?
1 Like
Also @BenMS suggested** elsewhere or above a while ago that when a SAFEcoin is used to purchase resources it is frozen. When the network uses up a SAFEcoinâs worth of resources then a frozen coin is finally deleted and available for farming payment.
This would limit the impact of people converting 100âs of thousands of SAFEcoins to âplay the marketâ and those 100âs of thousands are paid out in rewards. In the extreme case without freezing and allowing two way conversion we could see a lot more than the 4.3 billion coins worth of resources floating around the system as they are transferred for trading. But by Freezing coins the excess resource situation is very much limited and the problem goes away.
EDIT: ** it was only a discussion, not what SAFE does, but what could be done
2 Likes
THis reminds me of trading gift cards. Spent safecoin? need them back? weâll buy your unused safe network space 
4 Likes
Iâve edited the [original idea][1] based on input so far. I think itâs easiest to just call it what it is⌠âSAFE storageâ but added the (GB) suffix to let people know how much space it represents.
Thanks for everyoneâs input. This was really helpful.
[1]: Safecoin divisibility - #69 by dyamanaka
2 Likes
Canât divisibility be solved by using safecents instead of safecoins as the âgranularityâ? And instead of using 2^32 resolution to use 2^64 resolution.
100 safecents = 1 safecoin
And when 1 safecoin is transferred then under the hood itâs always transferred as 100 safecents.
1 Like
Every coin moved requires 28/32 consensus⌠If you use safecents then it will require 100 times the computing power from the network.
At least that is my understanding.
1 Like
One would expect that the network size would be large by the time we need cents or milli-sc and that there would be an abundance of groups all handling their own set of transactions. That is the increased workload is spread across a greater number of groups and in effect be similar node processing load to normal safecoins before the âcentsâ are needed.
Also in most cases (maybe not too) that you will only be transacting upto one safecoinâs worth of cents (or milli-sc) then the rest of the transaction will be safecoins.
1 Like
Then use tokens with fixed denominations:
1 safecent
10 safecent
1 safecoin
10 safecoin
100 safecoin
1000 safecoin
So for example when a user transfers 84,536.50 safecoins then the system uses 84 + 5 + 3 + 6 + 5 = 103 tokens.
Otherwise even without divisibility if a user transfers 100,000 safecoins then it will require 100,000 times the computing power! If thatâs really how the current implementation works.
And for good measure (in case the safecoin market cap goes to the moon), let the system from the beginning also have tokens for the following denominations:
10,000 safecoin
100,000 safecoin
1,000,000 safecoin
That will be useful when the international banksters start using the SAFE network.
Can I suggest, if you have not done so, read the whole thread and see the evolution of ideas. There has been a lot of valuable discussions in this thread and you can see one of the developers @BenMS who started the thread evolve his ideas on what can be done to successfully implement divisibility. It is really important that you have the background information in order to formulate more insightful suggestions.
It will save a lot a asking the same questions that have already been asked and discussed.
This was discussed too, and from those discussions you can make more informed suggestions. This is one suggestion that seems reasonable. Also suggested here or elsewhere was milli-sc (1/1000 of an sc)
It would also seem that there is still much to sort through on this subject and like Ben did before we need âdevâ input to help the discussion along with some expert help on what is easy or hard to do with the current code base and proposed logic
Ok, I have read some of the comments in this thread now. Seems to be a tricky problem. My solution is fairly simple. And a token for milli-sc can be included.
Would my idea be difficult to implement in practice? Iâm thinking that the system can automatically convert different tokens. So for example if a user transfers 99.99 safecoins and only has one 100 safecoin token in the wallet, then the system splits the 100 safecoin token into nine 10 safecoin tokens, nine 1 safecoin tokens, nine 10 safecent tokens and ten 1 safecent tokens. And then a transaction is made with a combination of those tokens.
And if a user wants to transfer 100 safecoins and has 100,000 milli-sc tokens in the wallet (because of having gathered micropayments), then the system converts all 100,000 milli-sc tokens into a single 100 safecoin token before doing the transaction.
Please read the rest, it was discussed at some length.
There are pros and cons to the involved and rather than repeat it all, just look for yourself. Hopefully it can be ironed out.
The over-riding principle behind any split is that there is only the designed approx 4.3 billion safecoin worth at any time. Having tokens adds complexity. So questions over which method will end up adding minimum complexity and minimise any network processing overheads.
[Suggesting changing the cap will not occur in the current development because of agreements with the people who bought in the pre-sale. Otherwise law suits would abound and SAFE would die]
Yes, even if a modification would be very beneficial for the IPO participants there is always a risk that someone would file a lawsuit anyway. Maidsafe could do what Donald Trump has talked about 
, first examine the contract very carefully and check whatâs legally possible and then make a win-win deal that is beneficial for both the SAFE network and for the IPO participants (who want a huge return on their investment because of the high risk).
Just a last idea: would it be possible for the wallets to do the token conversion themselves (by using some clever cryptography)? Then most of the complexity would be removed from the SAFE network itself.
When you read through, you also see the suggestion of an account based value that can hold say a safecoin/2^32 amount and safecoin converted (one way) into it so it can be used around the network and for transactions. Then have an exchange (eg safex app) to allow users to get back safecoin for an equivalent amount. The discussion progressed from there, but this sounds similar to your suggestion.
Thats why its very important to read the thread and understand the discussions, or else we will have post after post repeating the previous discussions. It is a very worth while discussion to have, just gets annoying repeating it After digesting the previous discussions then you can formulate further ideas/suggestions which hopefully leads us to a solution, and if not then we still get to expand our views of what might be.