Human App Ranking (App Mining 2.0 proposal)
sdsantos opened this issue · 14 comments
This proposal requires that all apps participating on App Mining are worthy and represent the Blockstack ethos, as proposed on #218
Once apps join the App Mining program, they would automatically receive a one-time reward of $500 to be awarded that month, similar to entering a "Can't Be Evil" competition. This would be an incentive for new developers to try out Blockstack, plus a light incentive for existing developers to experiment with new app ideas.
Then, every month, 3 entities (individuals or companies) would be selected to each pick their top 10 App Mining apps. In the beginning of the month, they would advertise what they are looking for in apps, and could require some extra info to be submitted by App developers. At the end of the month, each entity would annouce their TOP 10 apps and rewards would be paid out.
Suggestion for the Top 10 payouts:
- 1st place: $20K
- 2nd place: $10K
- 3rd place: $5K
- 4th-10th places (honourable mentions, no need to rank them): $1K
Selected entities should be pick to cover a range as wide as possible. Some will value usage and growth, others design, innovation, open-source contributions, marketing, etc... Each would have their chance to get to known Blockstack apps better, plus instil their values into the community.
Proposed strategy for finding entities each month:
- Pick Blockstack partners like the ones voting at Can't be evil competitions (dapp.com, Product Hunt, NIL...)
- One Blockstack evangelist picked by the PBC team, as a reward for his/her work (if he/she has any app on the program, it's not eligible)
- Hold a continuous voting poll where all community members propose their own entities. The PBC team would invite the entities in order until one accepts the responsibility.
Goals:
- Reward apps across a wide range of goals and verticals
- Reduce rank gaming as much as possible
- Expose more outside entities to the Blockstack ecosystem
Shortcomings:
- Not fully automatable
- The effort saved on collecting and auditing ranking data would be spent on inviting outsider entites to participate
Requirements:
- A higher bar for entering the App Mining program (#218). Not all entities are aware of Blockstack values, so we need to make sure that all eligible apps are worthy of being showcased.
- A better app directory than app.co for entities to find the kind of apps they are looking for. It would need information such as a longer description, screenshots, categories/tags, release notes, showcase external coverage...
I tried to make the proposal concrete but it's open to discussion, of course. Values like the number of entities per month or the actual monetary rewards come from a personal gut feeling, not an analysis of the amount of money put aside of the App Mining program, or game theory.
I do suggest that, if rewards become bigger than the suggested ones, winners need to redirect some of those earnings to specific bounties aimed at improving Blockstack tools and community.
Wonderful!
Comment/questions/suggestions spit out to enable discussion on each as an individual item.
Proposed amendment 1. I would like to see that number of voters up at 9-ish, minimum. All rotated each month. The number of voters would track along with app growth, voters = 9 or floor(# of apps/50), whichever is greater.
Proposed amendment 2. I'd like to see a commitment to a plan where the voters come from a DAO (Clarity smart contract) at some specified time in the future (July-ish, maybe?). I.e. STX holders will be able to vote for voters. Voters campaign on their voting record. (Dependent on STX v2).
Proposed amendment 3. I'd like to specify as part of the proposal that all votes are always transparent, via on-chain commitment. (Dependent on STX v2).
Proposed amendment 4. Related, all score calculations should be done via on-chain smart contract. Governance of changes to that contract would be owned by that same DAO.
Proposed amendment 5. I'd like to see payouts structures more along the lines of #219 , maybe something in between these two proposals because the structures don't exactly align.
Proposed amendment 6. I'd like to see those voters paid some minimum amount from the overall payout. This is no different than current state, where reviewers get some incentive to perform honestly (Try My UI, Awario, NIL, as well as Product Hunt and Democracy Earth before).
Proposed amendment 7. DAO should include term limits for voters (details TBD) to encourage some volatility and discourage stasis.
Proposed amendment 8. App Mining was never intended to provide continuous income, therefore apps would only get paid for a set amount of time, say 6 months max.
Proposed amendment 9. All payouts should be in STX only. Users with STX are always free to convert to whatever other payment type they like, btc, eth, or even legacy monetary systems. This is problematic today, but could be modified to enable an immediate switch to STX only after an SEC approved exchange lists STX.
Great Work @sdsantos
@dantrevino I agree with the proposal with your amendments. Can the proposal be updated with the amendments. Would make for easier reading.
Does this proposal anticipate maintaining current reviewers like NIL, TMUI and Awario? Or Will they be eliminated in favor of a single review panel?
Please note the provisions set forth in the Token Paper section 3 on App Mining:
"App reviewers have and will be relevant technology companies who have also proven the ability to develop or identify high-quality applications. App reviewers participate in a curation process by selecting eligible applications and assigning a normalized score to be aggregated into a final app ranking for rewards [8]. The app reviewers may also receive a percentage of the Stacks tokens distributed under the Reg A App Mining Program. Through the Reg A offering, $1,110,000 worth of Stacks tokens are reserved for distribution to app reviewers."
While I really like this: "One Blockstack evangelist picked by the PBC team, as a reward for his/her work (if he/she has any app on the program, it's not eligible)" it may violate SEC rules
based on Section 3 of the Token Paper Quoted above as well as the same language in the SEC filing.
In my mind, NIL, is a Gating Reviewer. ie, admission to App Mining is dependent on minimum score. The intention here is to provide a minimum set of requirements and standards before admission to app mining, similar to the reviews that Apple does for apps entering the app store. This needs further specification.
Final score calculation, in my amendment, would be specified in a smart contract and would include TMUI, Awario, and HAR.
Final score, per @sdsantos' proposal would only include HAR. I'll let him specify.
@joberding I have not taken in consideration that Section 3. If that's a strict requirement, then only outside companies/institutions could be invited to vote each month.
This proposal is starting from a blank slate.
- NIL would be needed as a gating reviewer like said in #218. And it could also be invited as an App Mining voting entity every other month.
- Awario has been gamed almost every month, so it can't be used a permanent reviewer. The Awario company could be invited as a voting entity every X months, for example. Or another voting entity might really value Marketing and Reach and use Awario as a tool to decide on their TOP 10 that month.
- TryMyUI is a really useful service, but biased towards simple apps for a mainstream audience. Again, the TryMyUI company could be invited as a voting entity every X months. Or another voting entity might really value usability and use TryMyUI tests as a tool to decide on their TOP 10 that month.
The main idea behind the proposal is really:
- The voting entities should vary as much as possible every month, so they are not predicable, and cover a wide range of different values.
- Each voting entity should award their own rewards independently, instead of combining votes into a final score like what happens right now, to ensure it's not only the lowest common denominator that gets rewarded.
You may think this proposal as a "Can't be evil" competition, like the 3 already organised so far, but happening every month and not only for new apps. I'm also ok with making it more automate, transparent and on-chain as much as possible, I'm just not an expert on the subject to propose how to do it.