Improvement of Dispenser Functionality

[davestaxcp]

Background:

the quick story of dispensers goes like this:

• DEX doesnt do BTC direct orders, which is what everyone always wanted when trading assets on counterparty natively
• BTC/[xcptokens] pairs required BTCpay
• nobody liked BTCpay, not many wallet providers offered it, and it was (is) clunky and confusing
• because XCP and other main tokens like PEPECASH and BITCRYSTALS were traded on cex's it wasnt too much of an issue
• cex's drop XCP and xcp related assets
• Villar proposes dispensers for BTC to xcp asset markets on a protocol level
• Years later dispensers become heavily used and relied on for all xcp projects and basically all liquid xcp supply
• users start to experiment with (and now rely on) 'multi-dispensers' (send 1 btc and get 50 separate assets) by broadcasting 50 different dispensers on one address for one price
• dispenser scams were noted during heavy use (no block delay in closing dispensers)
• 5 block delay added, most scam reports end (but still the possibility of front running the payment from the owner on the same BTC block)
• adding dispensers to 'non owned' addresses quirk introduced
• dispenser logic changed for only 'owned' unused addresses to be able to create dispensers
• only quirk that remains unsolved is the seller 'front running' dispensers on the same block as the buyer, but is seldom reported as happening
• people want an alternative due to seller trust issues (finding and contacting seller etc), PSBT and Atomic Swap proposed and implemented
• dispensers still heavily used after implementation to acquire XCP and other project assets even with risks
  (which is why Nifty does his verified xcp dispenser lists)
• users want to use dispensers for more complicated functionalities, sell whole wallets (see How will be the best way to sell a complete Counterparty address with multiple assets? #141), or randomized dispenses, or more importantly 'multi-dispensers' without costs associated

with the addition of adding atomic swaps, dispenser rules were changed under Adam

• only Counterparty wallets can initiate dispenser transactions (added new flag, no more using electrum or 3rd party wallets without the flag)
• dispenser creation on an unused new (but owned by dispenser creator) address was changed on a protocol level to not be possible to create in one tx (something used heavily and relied upon on FW desktop specifically)... from my understanding it was required by the protocol that the address have SOME btc history

... could be wrong here and FW could update to a new code structure, but because of arguments relating to dispenser changes, dev decided to take a back seat to putting time into FW Desktop.... furthermore since at the time this was one of the most used wallets, users simply said "they broke what i used to could do!"

• Horizon offers the 1tx dispenser opening on new unused address functionality but older wallets (and FW) do not implement this ^
• dispensers become 3x more expensive for the current "i dont want to use a new wallet" userbase (user have to send asset, then send btc, then send open dispenser tx)

---

Main Points:

• even with the frontrunning risk, dispensers are relied on and fully embraced in the counterparty community

1. • users want creation of dispensers to be cheap, when sat/vb increases, multiple dispenser creation is a headache and costly especially for users that prefer older wallets like FW desktop
2. • users want to create multi-dispensers (for example selling 100 assets for 1 btc) or their whole wallet asset balance
3. • users want randomized dispensers (see 'xcp slots' testing by Xcer https://xcer.co/free_rolls)

---

Purpose:

With new encoding methods Counterparty could implement new features for Dispensers:

• Cheap, 1 tx opening of dispenser on provable (... sign message?) owned fresh user btc addresses
• Cheap, 1 tx opening of multi-dispensers that offer not just one asset but multiple (opens the possibility to sell whole 'current' address balance
• Option to add random chance for specific assets in a multi-dispenser to be dispensed at set price

---

Notes on discussion:

I'm sure there are more points here, more history, more functionalities wanted with dispenser usage, more caveats on a protocol level and more ideas that can be discussed or implemented.

But know I'd like to put the past behind and focus on fixing/upgrading the user sided experience of dispensers.

I don't mind if some of the responsibility is wallet providers, but I do know on a protocol level alot of these issues can be fixed and upgraded for an easier ease of use for devs and users alike, especially with cheaper encoding techniques