A scalable way to mint arbitrarily large collections of NFTs that earn their owners a yield in STX.
- Mint an arbitrarily large collection of NFTs with a single transaction.
- The buyer, not the creator, pays the transaction fee for minting the NFT.
- NFTs generate passive income for their owner in STX if the owner locks them up to prevent them from being sold or accessed on-chain.
- Fire-and-forget minting. NFTs are instantiated on-chain via the smart contract on buyer demand. The NFT creator does not need to mint and sell NFTs once the collection is published as an NFTree.
- The NFT creator gets a commission fee each time an NFT is instantiated.
- Minting process rewards early NFT community members of successful NFT projects built as NFTrees.
An NFT is a representation of a binding between a principal and data on a blockchain. The principal is identified by a set of one or more public keys or smart contracts, and the data is a public arbitrary string of bytes (such as, but not limited to, an image). The binding is maintained by a blockchain, which provides a reasonable and programmatically-enforced assurance that only the bound principal can transfer the data to another principal.
NFTs are a fundamental building block for a user-owned Internet. Providing a way to prove that a particular piece of data is bound to a particular principal enables users to treat their data as captial assets within the context of NFT-aware applications. For example, an NFT can act as an API key to an application's features: only people who own a particular kind of NFT may access them. If the features are valuable -- as in, someone might be willing to pay for them -- then the NFT itself is valuable. As another example, an NFT can act as a row in an equity capitalization table of a company: the owner of the NFT is entitled to revenue from the company proportional to the amount of equity the NFT represents. As a third example, an NFT can represent user-generated data, such as artworks, songs, movies, novels, and so on, whereby the act of producing and selling NFTs that represent this data is the act of funding the user's creative process.
The act of storing the NFT's binding on a blockchain is the act of empowering individuals across the world to seamlessly treat their data as capital assets. This is what differentiates NFTs from similar monetizable principal/data binding systems like domain names and online game items. The open-access nature of NFTs not only enables any user anywhere to acquire, use, and transfer NFTs, but also produce whole new sets of NFTs with as-of-yet-unknown use-cases.
Regardless of how NFTs are used, they are usually produced in the same way:
-
They are produced as a collection. The creator instantiates NFTs in batches, such as to raise funds for a creative endeavor or promote the applications that use them.
-
Collections are produced in a series. An NFT creator may produce multiple batches over time, such as when they produce more data to sell or add more features to applications that use them.
This also applies to NFTs that are programmatically generated by a smart contract: depending on the mechanism, the NFTs minted this way can be treated as a single batch of NFTs minted over a long time period, or a series of 1-item NFTs minted over the same time period.
Once an NFT is produced, it can be traded via the blockchain to other principals for other crypto assets. This enables a market for NFTs from a particular collection or series to form, which in turn determines the spot price for an NFT in a particular collection in a particular series over time.
The state-of-the-art way to sell a collection of NFTs is for the NFT creator -- be it a person or smart contract -- to instantiate the bindings for each NFT in one of two ways: all up-front, or upon request from a buyer. Once instantiated, they can be transferred to other users at the creator's whim, such as by selling them. While straightforward, this introduces several problems:
-
It can clog the blockchain. If the NFT creator instantiates an NFT before there is a buyer, it needlessly consumes blockchain compute capacity.
-
It costs the creator tokens. In both approaches, the NFT creator ends up paying for the transaction fee in the underlying blockchain's tokens. While the creator can price the transaction fee into the sale cost, the current sale mechanism creates a recurring token cost that the seller must be willing to pay.
-
It creates a moral hazard for the NFT creator. The collection creator initially owns all NFTs, and effectively controls the market for them. The NFT creator could devalue their collection by dumping their holdings. Also, because the NFT creator receives payment for their sold NFTs up-front, they are in a position to exit-scam their users: they could promise to use the raised funds to build out an application for the NFTs, and then just take the money and leave.
-
It underutilizes and under-rewards the user community. Selling NFTs to would-be users this way only encourages them flip NFTs, instead of trying to make the project itself better. This is unsustainable, because nothing appreciates forever. If the NFT is sold as a way to raise funds to build out some good or service that the NFT will eventually be used to access, then this introduces a window of time where users would rather buy and flip NFTs instead of commit to the project long-term (distorting the market and creating a moral hazard for the NFT creators).
These problems are addressed by NFTrees.
An NFTree is an NFT that represents an arbitrarily-large collection of other NFTs and NFTrees, while behaving as a single NFT on the blockchain. A user instantiates an NFT off of an NFTree by submitting a proof that the NFT is represented in the NFTree, which then instantiates that NFT as a separate binding owned by that user. Once "plucked" from the NFTree, the NFT cannot be instantiated again -- it is exclusively owned by that user.
An NFT creator would mint a single NFTree for each collection they make. An NFT buyer would pay to instantiate an NFT off of the NFTree. This way, NFTs are only instantiated once there is demand for them, which prevents the blockchain from getting clogged.
Two of the problems with the state-of-the-art way of minting NFTs have to do with creating the right incentives for a successful NFT project. NFTrees address these two problems by introducing a mining protocol, modeled after how an arcade makes money.
The astute reader will have deduced that an NFTree is probably a Merkle tree of NFTs (and they would be right), and might be wondering how the NFT pricing mechanism works. The answer to this is that the NFTree commits to not only the NFTs, but also a number of tickets that they are worth.
An NFTree smart contract contains an internal "tickets" fungible token that users must first acquire and then burn in order to instantiate the NFT. This is analogous to how an arcade works: in order to win prizes (the NFTs) whose value is denominated in tickets, the user must pay to play arcade games that produce tickets. When they have enough tickets, they bring them to the prize counter and exchange them for the prize.
With NFTrees, users earn tickets by participating in a proof-of-transfer-lite (PoX-lite) protocol to mine tickets, much like how the CityCoins project works. They commit STX to the NFTree contract, and in each Stacks block, they win a fixed amount of tickets with probability proportional to how much STX they committed relative to everyone else. Once the user has earned enough tickets this way (or bought some from someone else), they can mint an NFT off of the NFTree.
The reason for introducing tickets for buying NFTs is that it gives the NFT creator a way to specify the relative worth of NFTs without specifying an absolute price. The ticket mining protocol ensures that the tickets/STX ratio is known at all Stacks blocks; from there, the price of each NFT in STX can be calculated. This partially removes the NFT creator's moral hazard: the NFT creator does not control the initial valuation of the NFTs; the users do. Moreover, that valuation changes over time based on user mining activity. Additionally, if the NFT creator wants to own any of the NFTs they produce via the NFTree, they must participate in ticket-mining as well.
To be clear, not all users need to participate in mining tickets. They can just buy them from someone who does. But even then, as will be explained below, the NFTree smart contract includes a set of marketplace functions that let users submit buy-offers for NFTs in STX (including ones that have not yet been plucked off of the NFTree), which can be fulfilled by other users who do have the requisite tickets.
The other reason for pricing NFTs in tickets is that it introduces a way for NFTs to generate passive income for the users that hold them, but do not sell them. Like the STX token, an NFT in an NFTree can be stacked -- it can be locked up so that it does not resolve on-chain and cannot be transferred for a time, but during this time, the user receives a fraction of the STX spent by ticket miners proportional to the ticket-denominated value of the NFT. For example, if there are 90 tickets-worth of NFTs stacked, and Alice stacks an NFT worth 10 tickets, then she would receive 10% of the STX paid into the smart contract for mining tickets over the duration of her NFT lock-up period.
The amount of STX users would receive depends on not only the ticket-denominated valuation of their NFTs, but also how much STX other users spend mining tickets. Since producing tickets is the act of mining them, the amount of STX flowing into the NFTree contract would equal the value of the tickets produced. Since the worth of the tickets is underpinned by the worth of the NFTs, the amount of STX a stacking user can expect to receive is determined by the current STX-denominated valuation of the NFTs yet to be claimed. In other words, users who stack NFTs are incentivized to expand the market for NFTs in the NFTree -- both by growing the userbase and making the NFTs more valuable -- since this is what earns them the most passive income.
Making it possible to stack NFTs addresses the third problem in the state-of-the-art: by giving the user a choice between using the NFT, selling it, or stacking it and earing passive income, the user has more of a reason to hold onto an NFT even if they can't use it for its intended purpose yet. If they stack it, they don't need to engage in a speculative market for buying/selling NFTs; they just need to get more people involved in the NFT project. This is ultimately achieved by making the NFT useful, which aligns the user's behavior with the long-term success of the NFT project.
This is not to say that speculation will not happen (it will); this is to say that NFTrees give users an alternative that offers positive cash-flow over time while the NFT project is being built out. This invests users in the long-term success of the NFT project.
As a nice side-effect of stacking, the market price of an NFT would include its discounted cash-flow on top of whatever fundamental value the NFT project offers. NFTs created from NFTrees would only become worthless if no one mines tickets for them.
There are five user rules in an NFTree project:
-
Creators, who mint whole collections of NFTs as NFTrees. They receive a commission each time an NFT is minted off of an NFTree that they created.
-
Ticket miners, who send STX to the NFTree contract to mine NFTree tickets which will be burnt to instantiate NFTs. Ticket miners can claim the NFTs for themselves, they can sell their tickets to other users, and they can atomically instantiate and then sell an NFT for STX.
-
NFT buyers, who spend STX to acquire NFTs by submitting buy orders to the contract. The contract holds their STX in escrow for the duration of the buy order's lifetime, which in turn provides a global insight into how much demand (in STX) there is for NFTs in the contract's NFTrees.
-
NFT sellers, who fulfill buy orders by either selling NFTs they own, or burning tickets to mint uninstantiated NFTs for buyers. In the latter case, a commission fee will be claimed by the NFTree creator for the sold NFT.
-
NFT stackers, who lock up their NFTs to accumulate a STX yield from ticket mining. The act of locking the NFT renders it unsellable, and unresolvable via other smart contracts.
Each type of user has a different set of APIs to use to carry out their roles.
There are two parts to this project: the NFTree smart contract, and the
nftree.js
command-line tool for making NFTrees. The smart contract is a
proof-of-concept, and should be tailored to your project needs.
Because NFTrees commit to both the NFT data and tickets, they are represented in
the contract as a (buff 64)
, which encodes three pieces of data:
- The NFT SHA512/256 hash
- The NFT's number of tickets
- The NFT data's size
The last field is included as an anti-DDoS measure for Gaia hubs and other data hosts that hold onto the NFT data, so they'll know how big each NFT is before trying to store it.
The encoding is as follows:
|--hash (32 bytes)--|--size (16 bytes)--|--tickets (16 bytes)--|
The size
and tickets
fields are big-endian.
Building an NFTree is the act of building an NFT descriptor that commits to
all of the NFTs in the collection. This is achieved by building a Merkle tree
out of the NFTs, and making the hash
field of the NFTree's NFT descriptor the
Merkle root.
Building the Merkle tree, the Merkle proofs, and NFT descriptors from a set of
NFT data is handled by the nftree.js
program.
$ cd ./src
$ ./nftree.js build /path/to/your/NFTs /path/to/NFTree/output
-
The
/path/to/your/NFTs
argument is a directory with all of your NFTs as files. In addition, there must be atickets.csv
file that has two columns:name
andtickets
. Thename
column is the file name, andtickets
is the number of tickets the NFT is worth. -
The
/path/to/NFTree/output
argument is a directory into which the NFTs will be copied, and into which NFT descriptors and NFT Merkle proofs will be written. Each NFT will be named after its SHA512/256 hash; each NFT descriptor will be named after its hash and a.desc
suffix; each Merkle proof will be named after its hash and a.proof
suffix.
The /path/to/your/NFTs
argument can contain nested subdirectories. Each
subdirectory represents NFTrees within the NFTree, and must have its own
tickets.csv
file for its files. An NFT descriptor will be created for each
subdirectory NFTree, such that once the NFTree is minted, the inner NFTs can
then be minted.
This program prints out the hex-encoded NFT descriptor for the NFTree as a JSON string.
For example:
$ ./nftree.js build /tmp/nftree-input/ /tmp/nftree-test
"8a480f3f87b03dc1d6b8270cd65fc0e77f7640d49f97d4cd3b789c9de2d280080000000000000000000000000000001f00000000000000000000000000000684"
In more detail:
$ find /tmp/nftree-input/
/tmp/nftree-input/
/tmp/nftree-input/foo
/tmp/nftree-input/foo/foo
/tmp/nftree-input/foo/tickets.csv
/tmp/nftree-input/foo/bar
/tmp/nftree-input/tickets.csv
/tmp/nftree-input/hello
/tmp/nftree-input/snarf
/tmp/nftree-input/boop
$
$ cat /tmp/nftree-input/tickets.csv
name,tickets
hello,123
boop,456
snarf,789
$
$ cat /tmp/nftree-input/foo/tickets.csv
name,tickets
foo,100
bar,200
$
$ ./nftree.js build /tmp/nftree-input/ /tmp/nftree-test
"8a480f3f87b03dc1d6b8270cd65fc0e77f7640d49f97d4cd3b789c9de2d280080000000000000000000000000000001f00000000000000000000000000000684"
$
$ find /tmp/nftree-test
/tmp/nftree-test/
/tmp/nftree-test/6ff3e7040fc45301764d3b5be9a01814a64f756545d869f4a44d9689e854bf1f.proof
/tmp/nftree-test/23000de7d22826a683fac628c926427ddd986913dba5332d6227085e746b577f.proof
/tmp/nftree-test/04c2f43e067d637611958ae92a54e90848dedeb5908bb3d3363cc57c573332b4
/tmp/nftree-test/23000de7d22826a683fac628c926427ddd986913dba5332d6227085e746b577f.desc
/tmp/nftree-test/d3dbb6686010b4b74742c52dfa8564f2e1501219568452a01b5a69d295d7d8c2.proof
/tmp/nftree-test/8a480f3f87b03dc1d6b8270cd65fc0e77f7640d49f97d4cd3b789c9de2d28008.desc
/tmp/nftree-test/332bc3d727592cdc62f40526cc2476d2627441656352b33da1eb53556c69cd17
/tmp/nftree-test/d3dbb6686010b4b74742c52dfa8564f2e1501219568452a01b5a69d295d7d8c2
/tmp/nftree-test/d3dbb6686010b4b74742c52dfa8564f2e1501219568452a01b5a69d295d7d8c2.desc
/tmp/nftree-test/6ff3e7040fc45301764d3b5be9a01814a64f756545d869f4a44d9689e854bf1f
/tmp/nftree-test/root
/tmp/nftree-test/bf73ee1fb7e8bf8fcdd5da06dd547052cd0f929a88f4aead68b218d3bde91134.proof
/tmp/nftree-test/6ff3e7040fc45301764d3b5be9a01814a64f756545d869f4a44d9689e854bf1f.desc
/tmp/nftree-test/332bc3d727592cdc62f40526cc2476d2627441656352b33da1eb53556c69cd17.proof
/tmp/nftree-test/bf73ee1fb7e8bf8fcdd5da06dd547052cd0f929a88f4aead68b218d3bde91134.desc
/tmp/nftree-test/04c2f43e067d637611958ae92a54e90848dedeb5908bb3d3363cc57c573332b4.proof
/tmp/nftree-test/bf73ee1fb7e8bf8fcdd5da06dd547052cd0f929a88f4aead68b218d3bde91134
/tmp/nftree-test/04c2f43e067d637611958ae92a54e90848dedeb5908bb3d3363cc57c573332b4.desc
/tmp/nftree-test/332bc3d727592cdc62f40526cc2476d2627441656352b33da1eb53556c69cd17.desc
The NFTree creator needs to upload the contents of /path/to/NFTree/output
to a
Web server, so they'll be available for viewing. The URL to each NFT will be
constructed by the NFTree smart contract by appending the SHA512/256 hash of the
NFT to a knwon URL prefix. So for example, if the NFTs are going to be
available at https://nftree-example.com/nft-data
, the NFT creator would put
the contents of this directory onto the Web server such that
https://nftree-example.com/nft-data/6ff3e7040fc45301764d3b5be9a01814a64f756545d869f4a44d9689e854bf1f
resolved to the file 6ff3e7040fc45301764d3b5be9a01814a64f756545d869f4a44d9689e854bf1f
.
The NFTree project creator can change the URL prefix by setting the
NFT_URL_PREFIX
constant in the nftree.clar
smart contract.
Once the NFT creator has the outputted NFT descriptor, then can call one of the
top-level functions to instantiate it on-chain -- (register-nftree)
and
(register-contract-nftree)
:
(define-public (register-nftree (nft-rec { tickets: uint, data-hash: (buff 32), size: uint }))
(define-public (register-contract-nftree (nft-rec { tickets: uint, data-hash: (buff 32), size: uint }))
The only difference between them is that the contract will own the root NFTree in the latter case, so no commission will be paid out for the root.
Both of these functions are used to mint the whole collection in /path/to/your/NFTs
. The NFT project leader(s)
would call this to instantiate new collections are they are made. All NFTs
represented in /path/to/your/NFTs
can then be instantiated by this smart
contract by interested users; they'd use the corresponding .proof
and .desc
files to construct the relevant contract-calls.
These functions return an integer NFT ID, which must be passed into functions
related to instantiating NFTs off of the NFTree (this argument is called
parent-nft-id
).
WARNING: If you make an NFT project, you will want to update these functions to limit who can produce NFTrees in your project this way. Usually, this will only be a designated admin, but any authentication rules you can write in Clarity are permitted. For example, you could have a DAO contract decide who can call this function.
To mine NFTree tickets, a miner can either commit STX in a single block, or over a range of blocks. The functions to do so are:
(define-public (mine-tickets (amount-ustx uint))
and
(define-public (mine-tickets-multi (amount-ustx-per-block uint) (num-blocks uint))
There will be one ticket winner per Stacks block, and the number of tickets granted to the winner is fixed regardless of how many or few STX are committed.
The winning ticket miner will not be known for 100 blocks, to ensure that the
commitments are sufficiently confirmed. At or after the 101st subsequent block,
the winner can claim their tickets by calling (claim-tickets)
:
(define-public (claim-tickets (miner principal) (blk uint))
Here, blk
is the Stacks block height at which the winner mined.
Miners can check if they were the winner in a block with (check-winner)
:
(define-read-only (check-winner (miner principal) (blk uint))
Users are not required to acquire and burn tickets in order to buy NFTs; it's only necessary that someone does this. Instead, users can simply offer STX for an NFT in the NFTree. Importantly, the user can offer STX for an NFT that is not yet instantiated, but is represented by the NFTree.
To submit a buy offer for a particular NFT, the user calls the (submit-buy-offer)
function:
(define-public (submit-buy-offer (nft-desc (buff 64)) (amount-ustx uint) (expires uint))
-
The
nft-desc
is the NFT descriptor for the NFT to buy -
The
amount-ustx
argument is the number of uSTX this buyer is willing to spend to get it. -
The
expires
argument is a future Stacks block height at which the buy offer expires.
When the user submits the buy offer, the contract escrows their STX so that the
NFT owner can later sell them the NFT. If the buy offer expires, then the buyer
can call (reclaim-buy-offer)
to get their STX back:
(define-public (reclaim-buy-offer (nft-desc (buff 64)))
They would pass the same nft-desc
as they did when the called
(submit-buy-offer)
.
A user can out-bid an existing buy offer by calling (submit-buy-offer)
with a
higher amount-ustx
offer for the same nft-desc
. If so, then the previous
buyer's STX are returned to them and the new buyer's buy offer replaces it.
The NFT in a buy offer does not need to exist in order to be fulfilled. If the
NFT already exists, then the NFT owner can sell the NFT by calling
(fulfill-buy-offer)
:
(define-public (fulfill-buy-offer (nft-desc (buff 64)))
They would pass the NFT descriptor for their NFT as the sole argument. If the call succeeds, then ownership of the NFT transfers to the buyer in the buy offer for this NFT, and the seller gets the STX escrowed by the smart contract.
If the NFT does not exist, then someone with tickets can fulfill the buy offer
by instantiating the NFT and then selling it to the buyer for the STX in one go via
the (fulfill-mine-order)
function:
(define-public (fulfill-mine-order
(nft-desc (buff 64))
(parent-nft-id uint)
(proof { hashes: (list 32 (buff 32)), index: uint })
)
-
The
nft-desc
argument is the NFT descriptor for the NFT to be instantiated and sold. It is stored in a.desc
file created by./nftree.js build
. -
The
parent-nft-id
argument is the NFT identifier of the NFTree that contains this NFT. -
The
proof
argument is a Merkle proof that links thenft-desc
to the NFTree identified byparent-nft-id
. Its JSON representation is created via a./nftree.js build
invocation, and is stored in a.proof
file.
If this function succeeds, the caller's tickets are burnt, the NFT is instantiated and then transferred to the buyer, and the caller gets the offered STX. The function returns the NFT ID for the instantiated NFT.
Once the user owns an NFT, they can stack it via the (stack-nft)
function:
(define-public (stack-nft (nft-id uint) (num-cycs uint))
-
The
nft-id
argument is the numeric NFT identifier, which is returned once the NFT is instantiated. -
The
num-cycs
argument is the number of reward cycles for which the NFT is stacked.
NFTs are stacked in fixed-length reward cycles, much like how PoX works in the Stacks blockchain. While the NFT is stacked, it cannot be transferred and it will not resolve via the SIP 009 interface. However, the owner of the NFT will receive a portion of the STX spent on mining tickets while it is locked up, proportional to the number of tickets the NFT is worth.
If an NFT is stacked, and the NFT is really an NFTree, then buy orders that instantiate NFTs off of it will fail.
To get the stacking rewards after the NFT unlocks, the user would call
(claim-stacking-rewards)
:
(define-public (claim-stacking-rewards (start-cyc uint) (num-cycs uint))
-
start-cyc
is the first reward cycle in which the NFT was stacked -
num-cycs
is the number of cycles in which the NFT was stacked
To determine what the start-cyc
is, the user can call the function
(blk-to-cyc)
to convert the Stacks block height at which they stacked their
NFT into the reward cycle in which it resides. The start-cyc
value is the
next reward cycle -- as with STX in PoX, the start reward cycle is the next
whole reward cycle at the point in time when the NFT gets stacked.
Each NFTree can represent up to about 4.1 billion (2^32 - 1) NFTs. If for some reason this isn't enough, an NFT in an NFTree can be another NFTree. In this case, the NFT hash field of its NFT descriptor is the Merkle root of the NFTs it represents.
If someone instantiates an NFT that happens to be an NFTree, anyone can then
proceed to instantiate the NFTs it represents by passing the NFTree's
parent-nft-id
value to whatever function is doing the instantiation.
Nested NFTrees are an advanced feature. They are meant to enable use-cases such as, but not limited to, the following:
-
Selling a collection of NFTs to a single buyer in one go.
-
Blocking the release of subsequent NFTs until enough tickets have been mined.
-
Implementing a semi-fungible token.
In all cases, the owner of the NFTree receives the commission fees for minting NFTs off of it. What this means is that if Alice registers an NFTree that contains another NFTree, and Bob buys the inner NFTree, then the commission fees instantiated off of Bob's NFTree will go to Bob, not Alice.
The proof-of-concept nftree.clar
contract implements SIP 009 faithfully, and
implements all but the (transfer)
function of SIP 010 for tickets. The reason for this
omission is because both SIPs have a (transfer)
function. In place of a
(transfer)
function for tickets, this contract offers (transfer-tickets)
with the same semantics.
The URL of an NFT is calculated by appending the hash of the NFT from its NFT descriptor to a URL prefix.
You will need clarity-cli
in your $PATH
. You can get it from
https://github.com/blockstack/stacks-blockchain.
$ cd ./contracts/tests && ./run-tests.sh
This repository is meant to be a proof-of-concept. In keeping with the Stacks Foundation ethos, this project is meant to set up other Stacks ecosystem participants to succeed. I will not be monetizing or productizing it for as long as I work for the Stacks Foundation.
As such, I encourage you to fork this repo and make your own changes instead of trying to send PRs.
Here are a few ways you can use this contract. A few modifications may be necessary, depending on the application, but they are straight-forward.
Imagine an application that shows high-quality content on some topic. NFTree creators act as journalists -- they gather and produce a firehose of content, and every so often, they register an NFTree with their new content.
- Small tweak: The contract will need to charge NFTree creators a registration fee, in NFTree tickets, for registering a new NFTree. The fee should be dynamic -- it should increase if there are more NFTrees being created, and decrease if there are less.
NFTree creators are compensated via commission fees whenever someone buys an NFT off of their NFTrees. Not all content they produce will be purchased, so they are incentivized to only produce content that will likely be bought (i.e. content that is on-topic).
A piece of content only shows up in the app if its NFT is actually purchased. The NFT buyers act as content moderators. The set of NFTs is visible to everyone, but it's a firehose of data -- buyers act as curators for the data, and in doing so provide a good experience for the app's users. Buyers are compensated by stacking their NFTs.
- Small tweak: The contract will need to be altered so that an NFT can only be stacked once, and within a fixed window of time after purchase. This is to ensure that there won't be free-riders who stack or accumulate lots of NFTs but don't provide any useful service in exchange for the income they'd receive.
Miners make money by mining tickets which are then sold to NFTree creators to pay their registration fees. In addition, they make money when they fulfill a mine order -- they receive the non-commission portion of the buyer's STX.
Everyone makes more money if more users are drawn into using the app. Users are drawn to the app by more and better content, which creates demand for both content creators and content curators (buyers), which in turn creates demand for miners.
A subset of users are incentivized to curate high-quality content, since if they draw more users in, they can recoup their purchase price by stacking the NFT later. Because they can only stack it once, and because the NFT can expire, they are incentivized to increase the amount of STX flowing into the contract before they stack and do it quickly. This incentivizes curators to grow demand for miner tickets, which is driven both by curation and by content creation.
Some users are also incentivized to become content creators, because high-quality content earns them a commission fee. The registration fee they pay in NFTree tickets drives demand for tickets, and thus income to NFT stackers. As long as the registration fee is lower than the expected NFT purchase price, creating high-quality content will be profitable. By making it so that the registration fee grows or shrinks with NFTree creation rates, there will always be a time when the registration fee will be low enough for a user to participate as long as the expected NFT purchase exceeds the blockchain transaction fee.
The system reaches a steady state when the profit margin for selling tickets equals the total cost for mining them (transaction fee plus STX committed), when the profit margin for stacking an NFT equals the cost of buying it (including transaction fees), and when the profit margin for creating and selling NFTs via an NFTree equals the ticket fee and transaction fee for registering them.
-
If creating NFTrees and selling NFTs becomes unprofitable, the registration fee will decrease until it becomes break-even or profitable again.
-
If buying and stacking NFTs becomes unprofitable, fewer buyers will stack them, causing stacking yields to increase, and thereby encouraging more stacking.
-
If mining tickets becomes unprofitable, fewer miners will participate, meaning that it will take fewer STX to win the same amount of tickets. As long as the expected ticket reward meets or exceeds the underlying blockchain transaction fee, there will always be at least one miner producing tickets for NFTree creators and buyers to purchase.
Imagine an application that acts like Medium or Substack, in which only free articles are available to the public, but subscribers can see additional content for a fee. NFTrees can implement a variant of this whereby a piece of content becomes public if enough people sponsor it.
To do so, an author breaks their document into a set of N Shamir Secrets, each of which is represented as an NFT. They publish an NFTree out of these shares, but they do not make the shares public. Once someone buys an NFT from the NFTree, the author discloses the share to the buyer. The buyer can then disclose the share if they wish. Once M <= N shares have been bought and dislosed, anyone can recombine the M shares and recover the original content.
- Small tweak: The contract will need to be altered so that an NFT's stacking power begins to diminish with age (but never goes to zero). This causes older NFTs to become less valuable, but it also means that early subscribers still stand to gain more STX than late subscribers. By causing older NFTs to depreciate, it "makes room" for new subscribers to join in the system (and it also encourages OG subscribers to keep subscribing to remain competative).
Buyers are like subscribers here, but they also share in the upside of the author if the author becomes popular enough to have many subscribers. Specifically, early subscribers gain more upside than late subscribers, since early subscribers can stack their NFTs and earn STX yield from them. This aligns subscribers' interests with the author -- subscribers are incentivized to help the author find more subscribers.
Miners gain from this because they must mine and sell the tickets to the buyers in order to mint the NFTs.
The author of course gains from having more buyers because the earn a commission on each Shamir Secret NFT they can sell.
Imagine a Web host that would keep hosting a website's content as long as somebody paid the bills. It doesn't have to be the website creator; in fact, the website creator can just walk away from the website and let its users take it over this way.
This can be enabled with NFTrees as follows. Every billing cycle, the Web host
creates an NFTree out of the hosted content and sets a price for hosting it for
the next billing cycle. All NFTs that are bought off of the NFTree remain
online for the next cycle; all that are not bought are removed from storage.
The tickets
field would reflect a meaningful measurement of the relative cost
to keep the data online.
The website itself would encourage visitors to buy NFTs periodically, such as by giving flares to sponsoring users or giving them access to extra features or content. In the limit, the website could structure itself like shareware -- only people who help pay to keep it going get access to the full features; everyone else gets only limited features.
By making it so that frequent sponsors can stack their NFTs, early sponsors can profit by helping make the website more and more popular as there is increased demand for NFTs (and tickets). This in turn crowdsources the task of helping increase the website's popularity -- as the website gets more popular, there would likely become more content to host, which in turn drives more demand for mining tickets, which in turn increases the yield on stacked NFTs.
Miners continue to mine and burn tickets to sell NFTs as before, but with the following tweak:
- Small tweak: To ensure that the hosting bill is paid in full, the
contract must treat the
tickets
field as the number of STX to pay, not tickets. To ensure this, ticket mining must be tweaked so that the number of tickets minted is equal to the number of STX put into the contract, instead of fixed.
This tweak gives users the ability to leverage the appreciation of STX to lower their hosting bill. If they commit STX at a lower price than it is when the hosting bill is paid, then they will have saved money, since the appreciation of STX in the mean time pays for part of the hosting cost.
The system works much like how crowdsourcing anything else works -- if enough people pay for a good or service, the service continues. The key difference is that there does not need to be a point person for gathering the money to pay the operating bills.
Once the size of the userbase reaches its peak will the act of buying and stacking NFTs become break-even -- the stacking revenue will match the cost of hosting the data, plus any markup the host applies, plus transaction fees on the underlying blockchain. But as the userbase is growing, stacking is profitable, since there is increasing demand for NFTs.
If the userbase shrinks, then the web host would delete unpaid-for site data. This in turn prunes the website of content that users no longer want, until maintaining the website becomes break-even with the new userbase size.