You can end up saying anything when you get things out of context. That note system idea (only a very old idea) was scraped for a much better balance system. And that note idea did not do as you claimed - oh well.
So carry on you seem to really be getting away from the SAFE network into a new network. So should I put this topic into related projects?
Yes, we are in full agreement on this point. I think that much of the discussion and/or disagreement on recent topics between ourselves has stemmed from certain things taken out of context at one point or another. Perhaps such is the difficulty with using a forum as the primary means of communication.
For example, consider the criticism and attack on rudimentary reasoning skills I received from @riddim and @oetyng for not presenting some basic “maths” about divisibility, reason, and scale before tossing ideas around of other more complicated or contrarian methods in the safecoin compo-cannibalism thread. What they perhaps failed to realize and took out of context was the whole premise which that thread had started from. The line of reasoning was based on what I thought was your criteria (perhaps again something inferred out of context) that scale and divisibility were rather limited resources due to network constraints; that it was a firm requirement to stay within 2^32 network objects to represent all safecoin currency regardless of the degree of divisibility chosen. It was this same criteria that led to my support for @norimi’s denominations concept since it is essentially a compression mechanism that allows one to have divisibility of similar limiting magnitude as that of 64bit addressing (meaning fixed point integers with 32bits for fractional part) but via a standard set of 2^32 objects. I was of the opinion that this fit rather elegantly with Mr. Irvine’s original description of having a total number of safecoin being no greater than 2^32, with divisibility being 1/(2^32), while also providing some additional flexibility as to the incrementation through the use of a scale factor in order to allow for rounded units such as 10^-6 SC instead of 1/(2^20) SC. This is just one possible concept in addition to the ones you have proposed, and others. The recent discussion in this forum just prior to my joining indicated that divisibility was still an open topic. I understood that to mean that your detailed proposal from nearly a year ago must have been long disregarded for some reason and the floor was open to toss some other ideas around.
As a newcomer to a project like SAFE, one attempts to digest as much information (some of which is constantly evolving) as one is able to link together from varied sources in order to “get up to speed”. In general I’ve always seen project forums as a place to try and bounce some ideas around, see what sticks, and what doesn’t. This allows for others who are more experienced to step in and help guide the knowledge transfer and identify a project’s communication styles in order to be productive, while also minimizing the echo chamber effect. It is not my intention to go off on unproductive tangents, nor do I want to, nor do I desire to try and steer the ship in any other direction than where it is already going. I can only ask that you forgive my ignorance on certain SAFE details and share links or references to information that might help me climb aboard the ship and help it move faster. Sincerely, if this is not the communication protocol for this thread/forum please let me know what is so that we can eliminate further friction and be more productive.
@riddim has certainly not attacked your reasoning skills. He has previously just asked you to respond to the specific figures. (Which you still seem to have avoided). And now he seemed to wonder why you, in the representation of the discussion made here above, leave out perfectly good reasoning which suggested that the idea for solution was unnecessary, since there was no problem to start with. And I agree, why would you do that? Either you confirm or refute it, but you don’t just ignore it. IMO. But hey, you go about it however you want
I on the other hand, have said that the problem was handled backwards, and in a later post that I was surprised by how you approached the problem. But that is not exactly attacking your reasoning skills. That is basically stating facts about how you went about it at this particular time, and then expressing that it surprised me (considering your profession, or learning).
The criteria you mention doesn’t really change neither how it was approached, nor the benefits of approaching it that way.
Ah well, all that aside. Keep on solving problems, I’m glad to see you here and giving it a go
One issue with your table is that because of the ico, you can’t guarantee that all current maid owner’s get their required allocation. It might work out fine that they can, and it might be highly likely that you can satisfy it, but not guaranteed. For example consider if no individual owned more than 5 coins, those 10k denominations would be no good. In order to guarantee the ico you would need about ~430MSC in 1SC denomination coins only.
1-2-4-8-… is good because regular: The divisors are always 2.
1-2-5 is “better” because: a) Just like normal money. Familiar to humans. Tech minds often forget how important. b) Close to regular: The divisors are almost always 2, except third of cases when it’s 1/8th bigger: 2, 2.5, 2, and repeat.
However. Maybe you are right. It’s already not 100% regular. But it’s regularly irregular. Similar could be implemented to tweak exponent for the “total value of coins in this denomination” series which is the 4th column on the picture. That column / coin denomination = number of coins in the denomination.
However. Tweaking is not so simple as you say. Like “4:8:5”. No:
First: Need to know distribution of transaction amounts. Log-Normal maybe good first approximation. Take Log-Mean as 0: It changes when Safe Coin price changes. So: Any number is good. Log-Variance: Take it from real transaction amounts. I don’t know where to get this data.
Then: Needs simulation to determine best locally irregular composition of coin denominations to represent division of market cap distributed as determined in first step. (Sorry. Complicated sentence. Took me time.)
“Locally irregular tweaking”: 3 steps, almost the same, juts little different. But every 3rd step is the same. I have this for denomination step: 2, 2.5, 2, 2, 2.5, 2, 2, 2.5, 2… For “series value” it’s about 1.468466, repeating. Tweaked, it could be: 1.468466 * a, 1.468466, 1.468466 / a, 1.468466 * a, 1.468466, 1.468466 / a, 1.468466 * a, 1.468466, 1.468466 / a… Globally, it’s 1.468466. But locally irregular.
Hey @norimi, I like your thought process and think you’ve covered some creative solutions. For example, I previously said :
However, I think I’m leaning towards team @neo now after looking through all the details including those on mutable and immutable data. You should check it out his proposal and give some critique in his thread. Your recent posts have me thinking more about divisibility lately, and I think a few different hybrid approaches might lead to the best of all worlds. When I have more time, I’ll try to throw out a well-researched idea for critique but in the mean time, if you get a chance please check out Is quasi-infinite divisibility useful? and let me know what you think in that thread.
I’m familiar with neo’s idea; it isn’t bad but it’s not to my taste. I fixed up my own. It brings truly infinite “divisibility” but without division and floating points. It self-adjusts to the level of adoption and it no longer has any kind of “lottery effect”. It has a limited market cap just as before. This means there is no inflation in spite of a potentially infinite supply of coins. It also comes with a constant farming rate so that small farmers can get regular rewards to avoid the bitcoin mining centralization problem. I’ll write it up soon.
Divisibility is a very popular topic on this forum to many people for multiple reasons. Off-topic questions in another thread made me decide to post some notes. I thought that we should try to consolidate the variety of ideas that are sprinkled throughout the forums and bring them here in “Safecoin divisibility” to reinforce it as the central divisibility discussion. Many of the links to the topics I covered are shown just under the original post, but they are easy to gloss over.
The summaries are brusk and not much by way of a literature review but it might help stop us from rehashing the same topics over and over. Reading this main Safecoin divisibility thread is also a treasure trove. The process of going through this in one sitting has helped clear up some things for me, and made some new connections. If this is just too much text, or if you think I wrote something in error, let me know and I’ll edit the post. Maybe these should be in their own thread, I’ll let the mods decide. The “highlighted Pros” or “highlighted Cons” tags refer to statements that were either for or against the ideas or topic introduced by the OP. The year tags are just approximate based on when a topic was initiated. Some threads concluded within the same year, but others (such as this thread) span multiple years.
If there is demand, I’ll post a similar set of notes for 2016 and 2017 in addition to this main thread as well in order to lessen the chance that people will gloss over the 313 posts that are now here. After that is done, we’ll likely need a summary of the summaries. It is interesting to see how many times similar ideas and concerns were “rediscovered”.
**TL;DR - SafeCoin Divisibility Notes : Part 1 (~2014)
Division in SafeCoin Summary: digipl described his distaste for binary division of safecoin and his preference for decimal division and the issue that 2^32 safecoin doesn’t allow for every living human to own at least 1SC. qi_ma described how the original conceptualization for safecoin believed it unlikely that a lot of divisibility would actually be needed (due to recycling economics?) but that the core routines allowed for any if necessary.
Subdividing safecoin Summary: This discussed the confusion surrounding the maximum 2^248 safecoin divisions presented in the Project Safe whitepaper and dissatisfaction with binary subdivision vs. decimal subdivision.
Increasing the total supply of SafeCoin Summary: seneca described his dislike of large amounts of divisibility vs. higher numbers of total safecoin due to innate human difficulties for thinking in terms of fractions. He suggested moving the decimal in the first 280bits of safecoin address that was described in the whitepaper. Various drawbacks were sighted and the idea was seen as violating the ico. Other considerations were that 2^32 SC were more than sufficient for network launch due to the internal producer-network-consumer market dynamic or “recycling”. dirvine stepped in to get everybody on track, indicating that divisibility up to 2^32 was possible with the whitepaper approach, but that he agreed with seneca that fractional payments are not as psychologically satisfying. He also described that the process was unidirectional, once coins are split they will exist has unique objects and not likely to recombine but could if necessary. Although denominations were criticized by some, dyamanaka presented a wallet structure that would show the total number of safecoins and fractional safecoins owned by an individual. These were taken to be Safecoin 1SC, SafeDimes 0.10SC and SafePennies or SafeCents for 0.01SC which was further described by whiteoutmashups.
**TL;DR - SafeCoin Divisibility Notes : Part 2 (~2015)
Safecoin VS SAFE Storage Summary: dyamanaka forked from the current Safecoin divisibility thread to discuss the wallet PUT balance as its direct use as a parallel currency to Safecoin called “Safe Storage” or “Safe GB” in order to handle the divisibility issue. He hypothesized network survivability in the far future. neo pointed out most of his concerns are alleviated due to how the network algorithm is not fixed and 1SC is allowed to purchase variable PUTS. The conversation soon transitioned to a discussion about network sustainability due to the the accumulation of “garbage data” in the network and limitations with relying on a Moore type law for storage. dirvine addressed this issue through the use of archive nodes and other means. krnelson pointed out the aesthetics of having safecoin valuations that are close to unity.
Safecoin divisibility vs. inflation ideology Summary: betterthantrav looked at the number of safecoins per person based on the number of users on the network and asked if inflation could be a solution to divisibility. Anders argued that crypto/safecoin inflation needs to be balanced with fiat inflation. dyamanaka pointed provided use cases for inflation vs. divisibility through an analogy, while others discussed the similarity to stock splits.
Safecoin Cap at 4.3Billion? Summary: Once again the question divisibility vs. inflation for equal investor value was generally discussed in regards to the future of safecoin. The consensus was that this would be possible by once again “freezing” a safecoin and introducing a parallel currency at some fixed exchange ratio. The use of SI unit system type labels for different denominations were also discussed for the case of decimal subdivision (milliSC, microSC, nanoSC etc)
Number of safecoins must not be to few Summary: There is some confusion considering the initial quantity of safecoin chosen to be 2^32 rather than something greater with lesser divisibility. fergish mentions that safecoin will not be divisible from the start due to the computational burden it would place on an infant network. neo mentioned how the PUT balance also serves as a means of divisibility to make network use more convenient to users. krnelson provided a summary of some ideas that had been presented in this thread (Safecoin divisibility) up until that point in time. He pointed out that in order for safecoin to properly serve the network it must be able to serve as a unit of account (UOA), being able to be easily divided and recombined. He also noted that supporting parallel currency denominations on the network is also computationally burdensome and is unecessary if safecoin is design well. Last, he mentions that high levels of divisibility are already required due to IOT.
Is the Safecoin Economy Deflationary and would it be better with Inflation built in? Summary: happybeing raised the question of how people holding safecoin for speculation of rising fiat will affect the internal safecoin ecosystem. Infinite divisibility (approximately) is discussed as the ultimate solution to deflationary spirals while not penalizing or stealing value from savers. seneca raised the point that deflation is not usually an issue for asset backed currency. riddim mentions the idea that safecoin itself need not be divisible but an alt-coin like SafeCents could be introduced to take care of concerns, and other discussions follow that consider other alt-coins denominations at fixed ratios analogous to traditional precious metal ratios that could be traded on decentralized exchanges and the problems associated with that process. It also seemed as though there was some confusion between inflation/deflation regarding the internal safecoin ecosystem vs. how safecoin may be treated by external human markets. It also seemed like there was a tendency to have fluid definitions of inflation vs. deflation rather than a common definition.
Not sure of demand, but since there were no major objections from the first set of folks to glance at these summaries, here are parts 3 and 4. That’s all for a while.
**TL;DR - SafeCoin Divisibility Notes : Part 3 (~2016)
On creating SAFE alt-coins Summary: seneca described a means for creating utility coins on top of the SAFE network that was applicable to Project Decorum. The method introduces a type of distributed “mining” system where users directly create the alt-coin by proof of work according to predefined settings introduced by the alt-coin’s originator and the natural properties of SD objects.
Will SafeCoin Be Divisible? Summary: General discussion on the benefits of divisibility and different schemes to achieve it. neo suggested limiting coin divisibility to 64 bits of fixed point precision, and mentioned how a 4bit ‘nibble’ mask can be used to convert from binary divisibility to decimal divisibility. There was also some confusion about whether 1SC would be allowed to purchase a fixed amount of storage or variable storage, as well as a description of using an SD (deprecated?) to handle subcoin wallet balances. The use of the “SAFE GB” PUT ledger was also mentioned by dyamanaka.
Most Wanted and Unwanted Features Summary: betterthantrav mentioned his desire for the network to provide incentives to users to delete data. neo responded with a description of how if PUT balance was used to handle safecoin divisibility, an internal “stock market” effect could lead to network instabilities.
Idea for divisible Safecoins Summary: polpolrene proposed a credit system where a safecoin is frozen, and then 1SC credit is given to the owner to spend decimal fractions of as they see fit based on discussions previously had on the forum with regard to safecoin economics and PUT balance. Nimos also proposed multiple dominations like 1 SafeCoin = 100 SafeCent. mav highlighted the fact that: “4.3B safecoins was decided as a compromise between the convenience of having lots of units vs the computational cost to the network when transferring lots of coins.” neo described how “Divisibility has then to be able to provide a lot of divisibility without multiplying the transaction load” and also mentioned that problems with transaction load can be mitigated by controlling the degree of safecoin divisibility based on network size. The discussion then continued with regards to algorithms that the network would need in order to know when a new degree of divisibility should be increased, and whether or not divisibility should be linked to fiat values. Erwin asked about simply increasing the number of coins by moving the decimal place and neo described the problem with doing this here.
**TL;DR - SafeCoin Divisibility Notes : Part 4 (~2017)
Proposal for SAFEcoin division - read datastructure topic first - please discuss Summary: neo proposed his idea for introducing safecoin divisibility by way of a wallet balance datastructure. There appear to be subtle differences with the proposal of that introduced earlier by polpolrene. Any portion of a coin up to 10^-18 divisibility is possible to transfer from one wallet to another using a single transaction. Farmers can be paid for each GET served with minute fractions of a safecoin, rather than having a certain probability of earning an entire safecoin. Some edge effects and security issues/non-issues need further discussion/clarification.
Divisibility of the coin Summary: General discussion on safecoin divisibility and how it relates to farming rewards. polpolrene provided a set of 4 recommended reads when it comes to divisibility. drehb referred to a previous discussion about how divisibility will help farmers be paid more regularly. neo highlighted the fact that even without divisibility, PUT balance provides some flexibility.
Proposal: TransferValue - a transferable value system of SAFE
polpolrene proposed a method to create an general method for creating alt-coins within the SAFE network analogous to the OMNI layer for BTC. This approach would allow for the creation of lesser valued denominations with which to handle ultimate divisibility concerns within the network. The conversation transitioned to talk about transaction fees due to the datatypes used, but dirvine mentioned they could be minimized.
Yet Another Safecoin Divisibility idea (YASDI - Network has a wallet) Summary: wes presented an idea for giving the network it’s own wallet, and it was discussed how concept was similar to that presented previously by polpolrene in “Idea for divisible Safecoins”. neo described the reasons why he prefers a balance system to denominations. toa introduced the concept of multiple owners for a single coin as a means to achieving a crude but intuitive/natural binary divisibility.
Yet Another Coin Idea Summary: Tim87 proposes a series of denominations following the powers of 2, and that this allows for large and small transactions while retaining the concept of discrete coins.
Safecoin divisibility / SafeNetwork potential market cap Summary: piluso began a discussion that compared the degree of divisibility required to adequately account for real world fiat values. neo advocated for 18 decimal places. polpolrene advocated for freezing safecoins and issuing other units of measure up to the required divisibility. There was also confusion from others about safecoin object size and population size, which led to a discussion how brute force divisibility suffers from storage and performance constraints
SafeCoin divisibility by ranges Summary: semtexzv introduced the idea of treating a single safecoin as a range of “partcoin” addresses. Double spending and unique ownership would be checked against range overlap. neo counters that a balance method does essentially the same thing and is easier.
YASDI - Bitwise Denominations Summary: mav further expanded the idea about the use of safecoin denominations, without resorting to parallel currencies or exchanges, but by using the bitwise relative abundance of similar safecoin addresses for identifying a coin denomination. tfa expanded upon this by demonstrating a straightforward means for having 2^32 safecoin by using a total supply of 4.444E13 network objects which would require at least 46 bit addresses. His example met the ico requirement of maintaining a hard cap of 2^32 SC, while also allowing for enough denominations of 1SC to provide for all current MAID holders and MaidSafe shareholders. There was continued confusion in regard to increasing address bit depth, where people thought it would require a greater total number of coins, rather than just increasing divisibility. neo accurately pointed out that the true
divisibility of bitcoin is not 1 satoshi but rather 1500 satoshi (divisibility = 1/66666.6666667). (sidenote: Considering that there the safecoin hard cap is ~204x that of BTC, this shows that 1/1000 of a SC or 1mSC provides approximately 3x the divisibility of classic BTC assuming worst case scenario if ALL coins are “hoarded”. As it has been noted in other sources, the effective divisibility gets even better to the network PUT/GET farming algorithms and the general use of the SAFE network )
Really good piece of work here, @jlpell. We all thank you!
I’m thinking this sort of summary would be really good to duplicate for other topics as well. (Not suggesting you do it, but something for anyone with a passion for a topic, or wanting to really embrace an understanding of it, to do.