So let's discuss incentives.
I'm not convinced that being able to spend Safecoin on storage would provide much incentive. If others can store data on the network without paying for it, how much is that storage worth spending on? Zero.
What if we set an arbitrary limit on how much you can store for free? Well, users will create additional client/vault pairs as needed (or their software client will do it for them), and carry on storing anything for free. Still worth zero.
But do we really need to tie Safecoins and online storage cost anyway? Couldn't we have people mining Safecoin while mostly ignoring the underlying system, much like it is with Bitcoins and other crypto-currencies?
In this case Safecoin's value would become purely market-based. I like the idea of having (a part of) the app-derived revenue spent on buying Safecoins. This could help bootstrapping the currency by seeding the market, although this would only come after a while.
---
About Safecoin supply:
Bitcoin miners nowadays are mostly incentivized by the newly-minted Bitcoins they get when claiming a proof-of-work. The supply of Bitcoins is limited however, and eventually transaction fees will provide most or all of the incentive to miners. Well that's the theory anyway, some academics seem to think transaction fees won't work so well for Bitcoin.
I do not see however how transaction fees would work for Safecoin. Would you mind expanding on that David?
Also, I'm not so sure about the concept of 'recycling' Safecoins. How do you make this recycling go at just the right pace to keep Safecoin mining rewarding, knowing that mining rates and app revenue could be largely uncorrelated? Do we have to rely on a central authority, even the MaidSafe Foundation, to keep the Safecoin economy going?
---
So, maybe we've got miners doing their thing for the sake of getting Safecoins. Maybe Safecoins are sought-after like Bitcoins are today, and people are investing heavily in them, with custom rigs and dedicated hardware. Does this mean that MaidSafe works, as an online storage platform?
I'm still not sure about that. I mean, even with a sizable portion of the World's population mining, you'd still have finite storage capacity, while offering infinite storage capacity for free to everyone.
Would people play fair?
Would storage capacity grow faster than data got uploaded?
If think that as soon as the MaidSafe platform leaves the hobbyist realm, the answer to the first question becomes "some won't". There will be abuses. There will be people trying to store as much as they can by any means necessary.
So it all hinges on the second question, and frankly I don't have an answer. Maybe no one does. What happens when storage capacity gets used up?
- Best case scenario, mining can still happen but its online storage by-product becomes mostly useless.
- Worst case scenario, even mining becomes impossible because a little bit of mining-related information needs to be stored on the network. The whole system collapses.
In any case, the primary goal of providing decentralized online storage isn't met.
I would be more comfortable with a system that guaranteed that such a failure would not happen, based on maths baked into its very structure. That's why I proposed chunk TTL.
It's very possible however that the system can offer strong guarantees based on modeling the world around it (that is, how people will interact with it based on economic incentives, how storage capacity will evolve over time and so on), but that's a much harder proposition, and much more of a leap of faith in the end.
Cheers,
Christophe Aguettaz
There's a risk of digressing into politics and ideology from this point, but if MaidSafe is to maximize its potential I think it needs to be able to leap that hurdle and keep developing on all levels. IMO there's no problem with diverse opinions on this or anything else, all are good in the pot, and from that can come a better consensus, and no need for any particular belief system to dominate, not even mine :-)So, following my belief for the moment... it may be that MaidSafe advances by recognising and incorporating the lessons of evolution, in ways appropriate to its internal operations (functions) and it's external environment, better than competing systems. In which case, we need to consider all the elements I've just enumerated while forming and reviewing these ideas.
That's why I think getting the structure right matters so much. It's all about setting up the reward and punishment structure so that selfish behavior is also desirable behavior.
In the same way as bitcoin, so each transaction would require an additional part of a safecoin to complete.
In the actual proposal this is where the network can calculate the average network data, this is why there is a POR. Think of this as a token tied to space or resources, but internal to the network. This POR can be purchased by safecoin, the amount will increase per safecoin as we need more resources, but they decrease in value over time. So today a safecoin may be £1 and buy 1Gb (a POR) but soon a POR will not be 1Gb it may be 1Tb, but also a safecoin may be £10. A safecoin (part)will buy this POR at whatever the (internal to network) market will sell it for, this price should be cheap and will likely be less than the £1 paid in the past for a 1Gb. Hope that makes sense. This is the graph of resource cost dropping and value of safecoin increasing.
This is where the miners are encouraged to meet demand. Everyone with a desktop PC can be a miner, this is not like asic wars.
So it all hinges on the second question, and frankly I don't have an answer. Maybe no one does. What happens when storage capacity gets used up?
- Best case scenario, mining can still happen but its online storage by-product becomes mostly useless.I am not sure what this part means, can you elaborate a little.
In the same way as bitcoin, so each transaction would require an additional part of a safecoin to complete.
I still don't get it. With Bitcoin, miners get rewarded because when they broacast a new addition to the blockchain, this addition contains the transactions rewarding themselves (in new bitcoins and transaction fees).
But here transactions are not handled by miners are they? Who's getting the transaction fee? Or does it go to 'the network' in general by being recycled?
In the actual proposal this is where the network can calculate the average network data, this is why there is a POR. Think of this as a token tied to space or resources, but internal to the network. This POR can be purchased by safecoin, the amount will increase per safecoin as we need more resources, but they decrease in value over time. So today a safecoin may be £1 and buy 1Gb (a POR) but soon a POR will not be 1Gb it may be 1Tb, but also a safecoin may be £10. A safecoin (part)will buy this POR at whatever the (internal to network) market will sell it for, this price should be cheap and will likely be less than the £1 paid in the past for a 1Gb. Hope that makes sense. This is the graph of resource cost dropping and value of safecoin increasing.
Yes, I understand that part, but I'm not sure how this relates to recycling being handled by one central entity.
Also, you agreed above that storing a limited amount of data (and perhaps an unlimited amount although you have your doubts) is basically free. Doesn't that mean that POR is worth 0?
This is where the miners are encouraged to meet demand. Everyone with a desktop PC can be a miner, this is not like asic wars.
So you say that miners are rewarded for mining 'above average'. Is that a new idea? I don't see it in the whitepaper.
How would this work? By setting a vault's mining interval using |healthy_space - average| instead of healthy_space alone, or something like that?
So it all hinges on the second question, and frankly I don't have an answer. Maybe no one does. What happens when storage capacity gets used up?
- Best case scenario, mining can still happen but its online storage by-product becomes mostly useless.I am not sure what this part means, can you elaborate a little.
I was just figuring that we could make mining resilient to network space shortage by having special rules for mining-related data. For instance each mining vaults reserves a fraction of its available space for chunks describing transactions, so that these go through even if vaults are otherwise full.
That brings another question to my mind. How do DataManagers find usable Pmids?
I imagine they'd have to iterate through random nodes until they find one that declares some usable space.
Now what happens if, say, 90% of Pmids are full? That'd be a hell of a lot of iterations to find a place to store to. At some point, even before the network is saturated with data, it'll become saturated with book-keeping traffic between DataManagers and PmidManagers.
Or won't it? Have you analyzed of perhaps tested cases where the network gets space-stressed?
--
Proposal - All users get unlimited disk space and pay zero POR/safecoin for this. All safecoin is mined by miners who are heavily encouraged (algorithmically to mine above network average space). In this proposal miners only, provide the network resources.
Cons:1: This puts a lot of the weight on miners in terms of their responsibility to the network. This is probably not too bad.
2: The resources put on the network are limited to miners, which may be a smallish pool, if the mining is not required to get POR for individuals.
3: There may be vast discrepancies in the resources per user in this case.
4: Most importantly, as safecoin is not recycled via the POR then how do miners continue incentivisation after the safecoin is all mined.
######################################If we can remedy the cons I believe the Pro's are compelling. This is still proof of resource, but for mining alone.Some ideas:1: Allow everyone to store up to the network average and pay in safecoin or use POR for above that. (this though means a good bit of work for MaidManagers and brings back as much management as we would have saved, so less efficient). Interestingly though even with the current system this would work and still allow all the pro's. So miners still go above network average, but everyone gets to network average without POR or vault being required. We still use POR in background though. This may be the half way house step.
2: Work out other mechanisms to ensure miners stay incentiviseda: Transaction fees (I personally do not like this, but that's personal)b: Do not put a limit on safecoin, but do make it harder as time goes past (easy to implement). This is a deflationary measure though, but then, is that in itself wrong?c: Somehow have app developers who get revenue still pay a 1% fee or similar and most of this (or all of it) goes to purchasing safecoin for recycle. (I think there is something decent here if it would work). I feel with the community we are building this would self police.d: In 2006 when I did the perpetual coin design, it was in itself deflationary, i.e. it went to zero value over time, this was to prevent hoarding and encourage spending. It's too big a proposal for here, so I mention it as an adjunct.
The SAFE network has to remove the safecoin cap and allow miners to keep mining. If the value of safecoin is constantly increasing, then inflation should be manageable.
I think it is important that this problem is looked at from an economics perspective. The SAFE network will essentially be its own micro-economy, with good and bad actors, all trying to gain maximum benefit.
Whenever there is a free resource, you are going to have the free rider problem (aka tragedy of the commons). What will prevent people abusing this? I suspect people will not delete old/stale files, they will not optimise their data usage and there will be much space wasted as a result.
Moreover, if this free resource is to be funded through creating new Safecoins alone (given to miners), this will act as a tax on everyone (as inflation of the money supply behaves like a tax on savings). As this monetary inflation will soon be reflected in prices, the miners will need to spend this money quickly, to gain the most value from it. The same will be true for others holding the currency.
While I can understand the desire to give storage to as many as possible, I don't believe this is the best approach.
If you want people to use the network efficiently, then they need an incentive to only use what they need. Therefore, there has to be a cost to use the resource.
If you want miners to contribute resources, then they need an incentive to provide as much as economical feasible. Therefore, there has to be a fee to use the resource.
With both of the above in place, you have a virtuous circle. The costs are driven down through miners providing more for less, while the network is used as efficiently as possible.
Ultimately, this drive for efficiency will be what gives poor people access, as the resource will get cheaper and cheaper. Ofc, there is always charity too, but fundamentally, the aim should be for the market to drive down costs to give access to all in the long run.
I'm only an armchair free market economist, but I would seek advice from Austrian economists if possible. They may be able to help with the subtleties. If the economics of this are too far wrong, it will be unsustainable and that will be a *great* shame.
I'm only an armchair free market economist, but I would seek advice from Austrian economists if possible. They may be able to help with the subtleties. If the economics of this are too far wrong, it will be unsustainable and that will be a *great* shame.
Moreover, as the (other) David said above at the end of his post, there are a series of major strengths for SafeCoin. That SafeCoin is just the tip of the iceberg of what this technology can deliver, I find that thought extraordinary - on paper, SafeCoin is already far better than Bitcoin with its $8bn market cap. That's food for thought!
Many in the community feel a fixed cap is the only way, I see this as a false economy as things just inflate. When I was a baby, a billionaire was a made up word.
Following the economic approach... Ways to make SAFE widely accessible:- diverse payment mechanisms that favor the poor. Are there things are not onerous for the poor to provide but which the rich can't exploit our control? Time is one possible differentiator, as poor people may find it worthwhile exchanging their time for resources, while rich would see this as expensive compared to "money". I'm not sure how this could be applied.- making it hard for "rich" to corner resources, or currency, in ways that make the poor have to pay more. Various tools here, as a tax on holdings (wealth) for example.- diverse pricing (e.g. small amounts free or cheap, larger amounts increasingly expensive - would "tax" heavy users, the "rich" more than light users, who would tend to be poorer )But what about the ecosystem analogue? I seem to have failed, so far, to inspire any such ideas! :-)Mark--
My Services & Blog:Sent from mobile - apols for autoccorrrct & brevit ;-)
So yes, I'm getting back to TTL again. David (or anyone else on the MaidSafe team), I gather TTL has been discussed internally before. Could you please summarize where you stand on this? What made you decide against it?
A person uses SAFE, totally happy he has stored all his data and family photo's will etc. and is comfortable. Then he goes off on a world tour for two years away from the computer.
Thoughts on David Irvine's Proposal:
Goals:
- Create reliable, *free* access to the network.
- Abandon the idea of "payment" with drive space because mobile adoption is the key to mass adoption.
- Remove friction & incentivize the transition between server based storage to encrypted, decentralized storage.
- Create a market based incentive scheme that both encourages early adoption (short run), and persistence (sustainability - long run).
Proposal:
All safecoin is mined by miners who are heavily encouraged (algorithmically to mine above network average space). In this proposal miners only, provide the network resources.
Comments:
- What is the effect of using "above average" as a mining incentive?
- Does it create barriers to entry for small miners? Economies of scale already is a large advantage when purchasing / building drives.
- Will the result necessarily drive small miners from the market? (too costly to be above average.)
- Should the network incentivize decentralization (more smaller nodes) in some way? Does that increase the speed and resilience of the network? How important are small miners to the ecosystem?
- For example if maidsafe's incentive scheme results in few large miners, could the simultaneous failure of several of the large miners (obviously extreme situation) risk data loss, network interruption?
Brainstorming thoughts:
- In the long run three factors matter: total availability space, total stability of space, & network propagation.
- Could an idea analogous to 'coin age' in Bitcoin reward continuous supply of space (and punish downtime)?
- How about rewarding data relaying & network propagation as well as useful space provided?
- There is an inherent problem in supplying unlimited free access to a finite non-free resource.
- What about a differentiation between permanent data on the network and flux data. Flux data would be free, open, but not guaranteed to be maintained.
- What about linking to the scarcity of Bitcoin. If you can sign a Bitcoin address that contains coins you can access permanent network space proportional to how many coins are being signed. (Note you do not need to send the coins - this is a solution to scarcity not payment)
Just some thoughts.
P.S. This weekend I went to my fathers house and took about 10 books on C++. I've been procrastinating learning to code (was going into finance). But Maidsafe & Bitcoin are beautiful things and I must be apart of the crypto-decentralization revolution :) You guys are doing something great.
3: One of my favourites : This also allows people in less developed countries who cannot afford the resources to mine or purchase safecoin the ability to connect, consume, create and add to the vast pool of human knowledge at no cost. To me this changes everything!
--
You received this message because you are subscribed to the Google Groups "MaidSafe-Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to maidsafe-develop...@googlegroups.com.
To post to this group, send email to maidsafe-d...@googlegroups.com.
Visit this group at http://groups.google.com/group/maidsafe-development.
To view this discussion on the web, visit https://groups.google.com/d/msgid/maidsafe-development/CAGYUzrYzmAWWiuut9wm8uV_80ukJQ3%2BALmVNMPd26oPtxtp85Q%40mail.gmail.com.
Live chunks are frozen when nobody pays for their upkeep. Alternatively, clients could request storing frozen chunks directly, and pay nothing.
To make this work, clients pay for reads. Either by 'thawing' the chunks, turning them into live chunks by paying for their upkeep, or with one-time GET costs.
So long as the network is reasonably far from full, everything gets kept in the network.
At some point though, if the network get too full, a garbage collection process is triggered that remove a sufficient amount of old chunks. There could be a reasonable lower bound on chunk age to avoid unpleasant surprises.
This way, there's no need to conjecture about user behavior anymore. If miners keep the network growing with demand, everything is kept forever. If however the network approaches dangerous load levels that could threaten its existence, it heals in the least damaging way possible to users.
--
It is normal for people to upload stuff and forget about them, especially if storage is free. This is like "clutter" and we should have a "MAID" clean up the network after sometime. If people want their clutter to remain on the network, they can pay safecoin to reset its time or status
see you in a minute, Im actually late and missed the beginning :)
You need to get into the code and help with some of this.
--
You received this message because you are subscribed to the Google Groups "MaidSafe-Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to maidsafe-develop...@googlegroups.com.
To post to this group, send email to maidsafe-d...@googlegroups.com.
Visit this group at http://groups.google.com/group/maidsafe-development.
To view this discussion on the web, visit https://groups.google.com/d/msgid/maidsafe-development/CAD5_TQf6sN6TZh-WgojCSLtSuNnUK5RAA0VfUUrWD%3DP2F4goUw%40mail.gmail.com.
I'll probably have more technical questions in the future. Is this list the right place to ask them? So far discussions have been pretty high-level on the dev list, is that the appropriate place to ask code-related questions, or is there a better suited one?
I've been reading this thread and the whitepapers with interest.It is a noble aim to give everyone in the world access to free data storage, but I don't understand how that can work in practice. As I see it, the facts are that storage space has an associated cost in terms of hardware and energy. In order to offer free space therefore someone else must be paying for the hardware and energy. Maidsafe appears to be proposing that "miners" provide the network with storage, and they are compensated with safecoin, generated through proof-of-resource. In order to pay for their energy bills and hardware investment, the miners must exchange their safecoin for national currency.What is it that drives demand for safecoin and gives it a value?
You might ask the same thing about bitcoin, but with bitcoin there are now hundreds of stores and businesses and people willing to exchange goods and services for bitcoin. These network effects give bitcoin its value. safecoin doesn't have this in place, so initially at least the demand for safecoin will be driven by speculators hoping to sell later at a higher price. If there is no goods and services trade in safecoin then the speculation bubble will eventually burst and the people who bought safecoin at the peak of its value will lose money. If the value of safecoin drops below that required to maintain network storage then "miners" will leave the network. The end result of this is that some speculators (people who get in early) will make money from safecoin, some speculators (people who didn't get out in time) will lose money. The investors who lose money are the people who end up having paid for the "free" storage offered by the maidsafe system.
If safecoin does achieve a level of adoption and is traded for goods and services then the "free" storage on offer is paid for by holders of safecoin, i.e. the reward paid to miners creates inflation which is a stealth tax. Obviously the bitcoin block reward is also a stealth tax on network security, but if maidsafe is offering unlimited free storage then the mining reward must cover miners costs or the miners will cease to provide the network with their storage. If storage is unlimited then it follows that inflation in safecoin must also be unlimited. If inflation in safecoin is not unlimited (i.e. if there is a cap on coin issuance from proof-of-resource mining) then it cannot be profitable to provide the resource (disk space) required to serve free-for-all storage space. If it becomes unprofitable to provide storage to the network then the network will not scale, i.e. it will quickly become full and therefore not able to achieve the stated aims.
Clearly if there is too much inflation then that will hamper adoption of safecoin as a real currency, users will prefer to trade in bitcoin.
I may have misunderstood some of the proposals, if so perhaps someone can explain who it is that ends up paying for the free storage on offer?
This is a bit of a lunch time braindump, but I think that for this project to be sustainable it is necessary to charge end users for storage. If you charge them in safecoin then that provides a built in value underpinning safecoin. The question then is how can we make it easy for people to obtain safecoin?
- early adopter bonuses (e.g. unlimited or, larger free-for-life allowances for the first million users).I like the second option because: a) it encourages users to sign up which helps the critical start up phase, b) it is less risky than "free to everyone", but can be changed to "free to everyone" (to great media fanfare) at any point, achieving David's commendable goal, but only when we believe it is sustainable.
It is a noble aim to give everyone in the world access to free data storage, but I don't understand how that can work in practice. As I see it, the facts are that storage space has an associated cost in terms of hardware and energy. In order to offer free space therefore someone else must be paying for the hardware and energy. Maidsafe appears to be proposing that "miners" provide the network with storage, and they are compensated with safecoin, generated through proof-of-resource. In order to pay for their energy bills and hardware investment, the miners must exchange their safecoin for national currency.
What is it that drives demand for safecoin and gives it a value?
You might ask the same thing about bitcoin, but with bitcoin there are now hundreds of stores and businesses and people willing to exchange goods and services for bitcoin. These network effects give bitcoin its value. safecoin doesn't have this in place, so initially at least the demand for safecoin will be driven by speculators hoping to sell later at a higher price. If there is no goods and services trade in safecoin then the speculation bubble will eventually burst and the people who bought safecoin at the peak of its value will lose money. If the value of safecoin drops below that required to maintain network storage then "miners" will leave the network. The end result of this is that some speculators (people who get in early) will make money from safecoin, some speculators (people who didn't get out in time) will lose money. The investors who lose money are the people who end up having paid for the "free" storage offered by the maidsafe system.
If safecoin does achieve a level of adoption and is traded for goods and services then the "free" storage on offer is paid for by holders of safecoin, i.e. the reward paid to miners creates inflation which is a stealth tax. Obviously the bitcoin block reward is also a stealth tax on network security, but if maidsafe is offering unlimited free storage then the mining reward must cover miners costs or the miners will cease to provide the network with their storage. If storage is unlimited then it follows that inflation in safecoin must also be unlimited. If inflation in safecoin is not unlimited (i.e. if there is a cap on coin issuance from proof-of-resource mining) then it cannot be profitable to provide the resource (disk space) required to serve free-for-all storage space. If it becomes unprofitable to provide storage to the network then the network will not scale, i.e. it will quickly become full and therefore not able to achieve the stated aims.
Clearly if there is too much inflation then that will hamper adoption of safecoin as a real currency, users will prefer to trade in bitcoin.
This is a bit of a lunch time braindump, but I think that for this project to be sustainable it is necessary to charge end users for storage. If you charge them in safecoin then that provides a built in value underpinning safecoin. The question then is how can we make it easy for people to obtain safecoin?
To view this discussion on the web, visit https://groups.google.com/d/msgid/maidsafe-development/qk0cxu142tjgsg0sfmk7n8uc.1395930112183%40email.android.com.
--
Kevin Croombs
Wow you guys have been busy.
I've been away for a week and it took me two hours to get caught up on these two threads.In general, I support this new proposal. But I think there are two factors which have been under-discussed.1. There is a time factor to the real-world cost of data storage that I don't see accounted for (unless I've missed it). Basically, it costs money to run a vault non-stop. While it's collecting data it gets paid in safecoin but it's hard to see how far those payments will go towards offsetting the operating cost because that is based on the valuation of safecoin. So jump to the end state, my vault is full and can't accept any more chunks meaning it can't earn any more. But it still costs money to operate. So what is the incentive to keep it running? I think there needs to be some kind of a vault maintenance incentive. Otherwise why wouldn't I turn it off and start again.
2. If data (or the storage of it) is the new gold, then I think there is a big difference between public data and private data. Public data is stuff like Wikipedia, blogs and open source software. I think it is generally in the interest of society to keep this stuff open and perpetually available. Private (including shared) data is of no value to society because it can't be read by anyone other than the owner (or a small group of people). I think that this distinction is where the line should be drawn between free use vs. paid use. I think we want it to be free for people to store information as this is a reflection of our society. In my opinion we're in danger of loosing a lot of our recent recorded history because our "public" data is stored on private servers and will eventually get destroyed. I should however have to pay to store my private photos, music and memories, just as I have to pay for my physical storage locker downtown.
And I think there is a definite link between the two. There needs to be a temporal element to my data storage fees, meaning I need to pay a monthly or yearly fee. There are a couple of (to me) obvious reasons for needing this: first my private data is encrypted and of no use or value to anyone else. And, as David has stated, the network can't even tell if I still need it. So I need to tell the network that I still need it, by paying for it periodically. When I die and stop paying for my storage my data will eventually be discarded (and the junk in my storage locker will be auctioned off). The second need is to ensure the efficient usage of the network. The basic premise is that resource that appears to be free will be abused until it is no longer free (see Canadian health care for an example). A small fee, and it needs to be small, will incent users to be more careful of what they store. Note however that the fee needs to be small, Google is already competing with SAFE by offering huge amounts of storage for a small amount of money. I (we) realize that there is a huge security caveat with that offer, but the average person doesn't see or care about that. They see cheap storage.
One possible solution to the pay vaults for storage over time could be paying for data served, rather than stored. I'd be surprised if you guys hadn't considered that so I'm curious as to the reasoning to not use that metric.
On the topic of capping the number of safecoins, I can't understand why we would do that. In my limited view, if data is the resource then safecoin is "backed" by the amount public data stored on the SAFE network. The more public data stored and reliably available, the stronger and more stable the value of safecoin. I think capping the number of safecoin hides that relationship and creates other problems with artificial solutions (like running out safecoin and trying to recycle them). But I'm not an economist, so we might want another opinion.
To view this discussion on the web, visit https://groups.google.com/d/msgid/maidsafe-development/CAP8xkoSd0f9EDO5YO8kQKEQ62otZ4-TVOXw6A_qad6i8h_zGcg%40mail.gmail.com.
I don't see how giving MaidSafe resources away is going to solve any third-world problems. Anyone who has access to a computer can already get several GB of free storage from a number of companies. It seems to me that user interface hardware and communications infrastructure is what is most needed in these under-served communities.
One possible solution to the pay vaults for storage over time could be paying for data served, rather than stored. I'd be surprised if you guys hadn't considered that so I'm curious as to the reasoning to not use that metric.This has become more talked about recently. It's not as simple to measure (requires extra fields in the record holders of chunks, these are held in PmidManagers). Of course mining attempts could be carried out on get attempts (data served). Hmm may be onto something here Chris. I wonder if we use data stored as the rate of attempts to mine and data served as the actual mine attempt. This feels quite correct! Any thoughts anyone ?
On the topic of capping the number of safecoins, I can't understand why we would do that. In my limited view, if data is the resource then safecoin is "backed" by the amount public data stored on the SAFE network. The more public data stored and reliably available, the stronger and more stable the value of safecoin. I think capping the number of safecoin hides that relationship and creates other problems with artificial solutions (like running out safecoin and trying to recycle them). But I'm not an economist, so we might want another opinion.
The network could calculate an inflation based on number of active accounts I think. Then more users means more safecoin. If we expand population we automatically expand the safecoin base by that amount. Maybe this is also correct.
For my own clarity, I think it's important to differrentiate between cloud storage and the SAFE network.
To me, the SAFE network is not just for storage. Google has already out-gunned us in that market.
Anyone can self-encrypt their own files before uploading to their free cloud storage.
The SAFE network "can" become the New Internet 2.0 (big pill to swallow, I know)
One possible solution to the pay vaults for storage over time could be paying for data served, rather than stored. I'd be surprised if you guys hadn't considered that so I'm curious as to the reasoning to not use that metric.This has become more talked about recently. It's not as simple to measure (requires extra fields in the record holders of chunks, these are held in PmidManagers). Of course mining attempts could be carried out on get attempts (data served). Hmm may be onto something here Chris. I wonder if we use data stored as the rate of attempts to mine and data served as the actual mine attempt. This feels quite correct! Any thoughts anyone ?Yes, this sounds like a possible solution. If the vault is doing the work, they should get paid.
For my own clarity, I think it's important to differrentiate between cloud storage and the SAFE network.
To me, the SAFE network is not just for storage. Google has already out-gunned us in that market.
Anyone can self-encrypt their own files before uploading to their free cloud storage.
Anyway -- lots of rambling, would love to know your thoughts.
Examples online promote and reward social behaviors:- Facebook/twitter: likes, "shared by", retweets- Reddit/stackexchange: upvotes/downvotes/karma points/trophiesExamples offline:- Utility bills showing your usage compared to average users in your area.- more?If we can use these ideas within SAFE for different user types, and developers, it could be very helpful.
Old
unused data is a reality, and should be archived/cleaned. My solution was to prioritize public over private. But I support the project either way.
The reason I believe
storage will not exceed demand is because production supply follows demand. Example: 10 years ago, most software only required X amount of MB hard drive space. Now they require X amount of GB of space. Naturally, companies produced larger hard drives to meet that demand. Eventually, there will be apps requiring X amount of TB drive space, GB bandwith, and consume a lot more electricity. I hope I'm wrong about this trend. Maybe we will discover free energy?
Anything can happen :)
Great blog David.
Deep, deep, deep discussion on this one.
You covered several topics so I'll just give some highlights.
Old
unused data is a reality, and should be archived/cleaned. My solution was to prioritize public over private. But I support the project either way.
The reason I believe
storage will not exceed demand is because production supply follows demand. Example: 10 years ago, most software only required X amount of MB hard drive space. Now they require X amount of GB of space. Naturally, companies produced larger hard drives to meet that demand. Eventually, there will be apps requiring X amount of TB drive space, GB bandwith, and consume a lot more electricity. I hope I'm wrong about this trend. Maybe we will discover free energy?
Anything can happen :)
[snip]
> The reason I believe storage will not exceed demand is because production supply follows demand. Example: 10 years ago, most software only required X amount of MB hard drive space. Now they require X amount of GB of space. Naturally, companies produced larger hard drives to meet that demand. Eventually, there will be apps requiring X amount of TB drive space, GB bandwith, and consume a lot more electricity. I hope I'm wrong about this trend. Maybe we will discover free energy?
I've always seen this as the other way around. Back in the eighties I foolishly said to a much brighter software developer colleague "soon computers will have vastly more disk and memory space than we'll ever be able to use". He, quite correctly pointed out that software expands to use the resources available, and this may be one reason he ended up head of cloud services at Netflix, and me enjoying life off-grid grid on a boat! :-)
What he saw, and I immediately realised was correct, is that computing resources mean we can build more powerful tools, leveraging computer power to human productivity (e.g. higher level languages that generate less efficient code, but increase developer productivity).
Aside: After the Snowden backlash began towards cloud services I cheekily asked him on twitter, if he thought it would seriously damage commercial cloud services. He said no. I wonder who will be right on this one! :-)
Of course it isn't a simple either/or, about which follows what. It's both, so we need to consider things from both (all?) perspectives.
Message complete this time :-)
Mark
Old
unused data is a reality, and should be archived/cleaned. My solution was to prioritize public over private. But I support the project either way.
The reason I believe
storage will not exceed demand is because production supply follows demand. Example: 10 years ago, most software only required X amount of MB hard drive space. Now they require X amount of GB of space. Naturally, companies produced larger hard drives to meet that demand. Eventually, there will be apps requiring X amount of TB drive space, GB bandwith, and consume a lot more electricity. I hope I'm wrong about this trend. Maybe we will discover free energy?
Anything can happen :)
The loop is incomplete - DavidI: you are not clear on where you see the value of SafeCoin arising - you mention developers having the option to charge. Is that your proposal? If so, we must ask if it is sufficient in the long term, and how we can bootstrap it.The will be some value simply through people's perception (as in speculators / as in bitcoin) at least to kick it off, but that is risky, and may not be sustainable, because other coins will no doubt be developed to provide highly optimised solutions to real world use cases, which will sideline coins which have no real world function, or are sufficiently less suitable. So there has to be something unique about SafeCoin IMO. What is it?
You said the needed to be a loop. My question is, how and on what, do people (miners, users, developers) spend SafeCoin, in order for it to get back into the network, and subsequently paid to miners?
Last email fired prematurely! This one completes...
I'm not against separating along public/private but not sure it makes sense. It ignores:- the immense value of free private secure storage for end users, which is value for the network since it will be nothing without mass appeal to *these* users. Solution: first xxxGB private data in free?
- the split between commercial and non-commercial public data. E.g. Wikipedia v Facebook, or Oracle's website etc. Do we want Oracle riding free while Nelson in Kalunda village pays to store his private woring notes on curing cancer?
I've always seen this as the other way around. Back in the eighties I foolishly said to a much brighter software developer colleague "soon computers will have vastly more disk and memory space than we'll ever be able to use". He, quite correctly pointed out that software expands to use the resources available, and this may be one reason he ended up head of cloud services at Netflix, and me enjoying life off-grid grid on a boat! :-)
Your friend is correct, about software expanding to use max resources available. However, the way production works is based on demand. Toyota Motor Corp makes X amount of cars each year, based on their potential sales calculation. They could easily produce more cars than is needed, but they don't. Same thing happens with hard drives. Companies could produce drives in excess of 4TB. But they won't until demand hits critical levels where it becomes profitable to do so. Just like Bitcoin with GPUs, when the SAFE network reaches adoption, there will be a surge in hard drive demand. Price levels will spike, and then you see many new hard drives hit the market with space beyond 4TB.
David,I feel unsatisfied by this - it's still vague in respect of what I'm after so would like us to try and clarify this.
You seem to be saying the ways SafeCoin can be given value are:a) in its own right, in the real world (similar to bitcoin), independent of its usefulness my relation to SAFE
b) as a fee (commission on app revenue) charged to developers of apps which generated revenue for the developer
c) as a fee charged for use of network services (to developers, to users of MaidSafe apps) - I'm not sure how this fits with "free for everyone", or if you have services in mind
d) for lots of things it will be easy for us to think up - well let's think some up if it's so easy, or it starts to look a bit like snake oil don't you think?
Are the above correct? Anything I've missed?
Do you see what I'm after? Broad brush is good up to a point, it's important, and particularly for brainstorming which I know is your focus just now. I am though concerned about the lack of concrete ideas for us to evaluate and refine in this critical aspect of the ecosystem.
David. Do you believe we can succeed without SafeCoin being the only way to buy certain services?If so, why!?
David. Do you believe we can succeed without SafeCoin being the only way to buy certain services?If so, why!?
OK, some examples, good. This is what I'm looking for. Without the, you won't have your tie in, and fate I painted awaits.This is so important that I think we need to have a credible set of examples which are feasible and consistent with the other things we're brainstorming.
[snip]> In essence safecoin should be the only way to purchase these services on the SAFE network, although exchanges will happen. The re-cycling is the killer app in this case and I think the search for that type of application is important.Agreed then, but this won't happen unless we make it happen, and if we rely on coercion the task will be unmanageable, so we need to at least make this the easiest way, if not the only, way to charge for SAFE apps.
> When speed, efficiency and anonymity are required then safecoin is great.Is this so? Is there an implementation spec or whitepaper?
On Wed, Apr 2, 2014 at 9:56 PM, Mark Hughes <mrh...@markhughes.com> wrote:
OK, some examples, good. This is what I'm looking for. Without the, you won't have your tie in, and fate I painted awaits.
This is so important that I think we need to have a credible set of examples which are feasible and consistent with the other things we're brainstorming.I think we can create several threads for each example, but I do not think we can foresee what developers will do and make available. I have some ideas, but this community is way more than me for sure. I am glad about that.
[snip]
> In essence safecoin should be the only way to purchase these services on the SAFE network, although exchanges will happen. The re-cycling is the killer app in this case and I think the search for that type of application is important.
Agreed then, but this won't happen unless we make it happen, and if we rely on coercion the task will be unmanageable, so we need to at least make this the easiest way, if not the only, way to charge for SAFE apps.I agree, I also think safecoin will be the only way as time goes by. This is where I think a very strong community will help, if the first app was a voting system then this would be much easier.
Not so much, I can try and explain more in another blog post. Essentially as safecoins are in themselves digital assets, then the transfer will be a network atomic digitally signed update to a coin. Each coin will have previous owner who signs the current owner. This means that to prove coins you will have to give the addresses to a person, they can confirm you own the coin. To transfer it you (the current owner) can become the previous owner and sign the new owner. In this way we preserve anonymity, but importantly the process of the signature si fast, therefor signing over a coin will be anonymised to reveal only the previous owner and fast as each coin (bit of data) is signed and stored (via a PUT request signed by the current owner). In that way we do not wait for a chain of hash proofs. The transaction is very fast and amortised across the network.
> When speed, efficiency and anonymity are required then safecoin is great.
Is this so? Is there an implementation spec or whitepaper?
Copy of a coin is useless as only connected nodes can confirm a coin by asking the network for it. This prevents double spend (why atomic transactions are important), which is the key issue in crypto currency. Any differences are easily resolved as each coin will have a transaction number (increment ++trans_num).
there is a good bit more., I will try and do a full paper on this. Essentially it is a very fast anonymous and simple operation. If a node is internet connected it can also tie in with the bitcoin blockchain allowing currency conversion between crypto currencies. It is likley this is the case for outlets who whish to confirm transacitons.
Each coin is splittable 10^8 and each part is contained in a single (in terms of address) network location, so change is very fast as well.
We will do a full and detailed paper, but think of safecoin as a digitally signed atomically updated piece of data. The type is specific to allow transaction handling rules are enforced, but apart from that it is as safe as data itself. This is the distributed block chain mechanism which is not simple as their is no chain :-) I hope that helps a bit, don't stop if it does not, it's important for everyone to fully get as much as possible regarding these things.
One possible solution to the pay vaults for storage over time could be paying for data served, rather than stored.
Just a quick thought related to this: If Vaults are paid to serve data then perhaps it would be benificial to have varying rates depending on how rare/old the data is. This could create a extra incentive to accommodate infrequently accessed data as the vault earns more to serve it.