Cardano Smart Voting

Intro

The number of proposals as well as their complexity is growing fast and it's becoming harder and more time consuming to vote on a proposal's merit and and one's personal interests. While reviewers do give "star" ratings, that help with quickly judging the overall quality of a proposal, a vote - especially when it's going more in the direction of governance - is always also a choice of personal preference and between trade-offs.

See a summary in this presentation: [ https://docs.google.com/presentation/d/1k7ggfI_sIO_A6N9ogUbHS2h3IlP2c1vxMXvJbCIixMU/ ]

Competing Interests and Trade-offs

Let's look at an example: While one user might be interested in proposals that are all about NFTs and Gaming, another user might be much more interested in API development. I think we can find a small number of such "competing interest descriptors" that can capture a good portion of the diversity of interest of voters in an easy format.

Following the example above we could call this "competing interest descriptor" the Entertainment descriptor. It has, for example, a range from -10 (boring) over 0 (neutral) to +10 (entertaining). It would be only a small overhead for reviewers to rate every proposal based on such a descriptor, similar to the current star-rating. In essence a tag with a numerical value.

 A number of different such "competing interest descriptors" come to mind: Whale-friendliness, Adoption-orientation, Local-Global, Scientific, DeFi*-oriented*... Now, I'm fully aware that such simple words don't capture the full depth and would provide ample space for differing interpretation and also the choice of number and quality of these descriptors itself is something that should be done carefully, which is why I think this should be a fluid community process, where community members can propose and improve descriptors and vote on their usefulness (which may change over time, as Cardano matures).

For example behind each of these descriptors there should be rationale, why it was chosen, what the extreme opposite ends of it represent, with examples etc. It is important the this representation is well balanced, with clearly laid out trade-offs. I imagine that each descriptor eventually becomes enriched with links, calculations  etc. that allows users to find their own position based on solid, objective information and to engage with others in constructive dialog about advantages and disadvantages of various positions.

Rather than compartmentalizing, this should allow users to orient their opinion in relation to others. For example an NFT/Art descriptor could make an argument that in the entire blockchain industry XX% of trades are for NFTs while it's less on Cardano, so this should be strengthened. The counterargument could be, that a solid base protocol without specialization will attract this automatically, so a special emphasis is not necessary.

Immediate Usefulness - Smart Voting

So what's in it for the voter? Let's assume this set of descriptors exists and we have, say 5 that do a reasonably good job to describe every individual proposal. Now all a voter has to do, is to locate his personal interests on these 5 dimensions to get an automated voting list, ranked by alignment with his interests right down to those that are opposed to his interests. 

So answering literally 5 statements with "disagree"-"fully agree" or dragging 5 sliders on a web interface (maybe complemented with filtering options) could give a user his personal voting lists in seconds. This would not diminish the voter's ability to dig deeper of course, or make different choices etc. 

(If voters have records of their voting history, this could also work the other way round - bases on past choices, the voters' personal interest descriptors could easily be calculated.)

Incentives & Business Model

While I see this more as a community driven effort, a well kept webpage, the data recording backend, moderation etc. will need a sustainable financing model. One way to achieve this, could be this model: via dApp integration users could earn a token or small amounts of ada by providing data for the system (descriptors for ideas for example). On the flip side, some advanced functionality (e.g. proposal ranking with more complex algorithms) could cost a small amount of the token or ada - maybe in the price range of the voting rewards. Balancing these for a small gain, could generate the income for keeping the system running long time. 

 So basically: Someone who puts information into the system (CAs, community members, ...) gets rewards, someone who uses that information (voter using an algorithm, interest groups looking for information about voter opinions, ...) pays for it. (In the case of voters maybe about as much as the voting rewards)

Roadmap, Costs, Stage-Gate etc

For the scope of this funding round I only do ask for initiation costs to follow up on the ideas presented here. That includes engaging with CAs and vCAs, finding scientists who published about VAAs (Voting Advice Applications) to map their experience into DAO scope, maybe commissioning a short report, estimate back- and frontend costs and to find a group of people wanting to start this - respectively integrating with one or more of the groups also working on Cardano Distributed decision making.

I'm fully aware that this might seem to be too much on the edge between idea and project, still, you got to start somewhere and winning a small funding, would go a long way to motivate to pursue this in earnest. I also do not see this as a business opportunity as such, that justifies personal investment (of time and other resources) but as the start of community effort for a new way to govern a system.

Weeks 4-6: Scout for team to support this ideas. For this I'd liase with CAs, vCAs but look around in genereal to find Cardanians with a skill set around back/front end programming, survey and data-science specialists and maybe marketing. In parallel work with other teams of this and previous funds and map out how and if it makes sense to integrate (for example: keep a separate webpage vs. integrating with an existing group) to maximize impact.

Weeks 6-10: Put together a plan with the team to outline all elements necessary to put things into practice. This includes cost and time estimations of database backend, frontend programming, server requirements, security measures, marketing costs etc. For some of these parts, payed estimates, or little studies might be necessary - for example to estimate user numbers, storage requirements, design a database scheme etc.

Weeks 10-12: Based on the previous phase and the most promising avenues, put out a fully budgeted and detailed proposal to fund the next stage. The next stage would probably target the functionality of finding and improving Opposing Interest Descriptors only and turn out simply ranked voting lists for individual users. Things like algorithms, DID, user and wallet integration, dApp connections and algorithm would come at a later stage.

The budget I ask is to cover any out-of-hand costs during these 12 weeks. Like buying a domain, little assessments, maybe a logo etc. and generally incentivize the time and effort collaborators (and myself) are investing to build this idea.

Looking further (beyond the scope of Fund 6)

It might seem overly complicated to find this descriptors via community choice, maybe on a dedicated homepage, with examples and arguments and information versus just sitting down a committee and churn out a simple list of such descriptors (though at the beginning this might actually be the solution of choice), so why am I proposing this?

Simply because such a mechanism could have much broader usefulness and implications in the future, so that setting up a robust mechanism might be worth the effort.

Imagine hundreds of thousands of voters voting on thousands of proposals (including Catalyst, but maybe also CIP and even Cardano network parameters). Currently Cardano docs say, that in the future, votes may be delegated to experienced and trusted community members who's interests align with one's own (not dissimilar to politicians). My proposal would lay the ground work for something new: algorithmic voting.

(Going a bit technical now, sorry.) The descriptors for anything that is put up for a vote, place this thing in a n-dimensional space. There are algorithms that can cluster such information and calculate the (n-dimensional) distance between the personal set of the 5 descriptors and single ideas or clusters of ideas in this n-dimensional space (see here for example (no affiliation): https://github.com/chriswernst/dbscan-python or https://en.wikipedia.org/wiki/Hierarchical_clustering ). Clusters in this space are nothing but interest groups - basically mathematically emerging parties. So instead of delegating their vote to a "pool operator politician" for voting, pool operators could have their very own algorithmic "Pool Party" (so to say).

A concrete example again, using the interest descriptors from above: There might be a cluster of "Gaming & NFT interested in local communities" or a cluster of "Developers into DEX and trading" etc. It is conceivable that such clusters and their operating parameters would allow users to delegate their vote to an on-chain algorithm or consult such an algorithm to assess even highly technical decisions that are put up for voting based on their personal interests. 

From a simple set of descriptors, a complex ecosystem of interest groups, parties and voting could emerge that would allow voters to find their place among competing interests, assess how their personal interests align with voting choices and hopefully also get a better understanding of opposing positions, leading to constructive governance that accurately reflects the interests of all participants.

How success looks like

For Fund 6: Found a team of engaged and knowledgeable individuals who want to build this. Have a solid battle plan out to make it happen via the next funding round.

For the next Fund: Have a own or integrated website that allows CAs / users to propose and improve Opposing Interest Descriptors and rate Ideascale Funding Ideas bases on these. Have a system in place that gives users a personalized, ranked voting list based on their interests after answering simple questions.

For the Fund after that: Integrate everything with dApp connection, DID and add algorithmic voting, automated interest group clustering and other advanced functions. Users will be able to connect their wallets and delegate votes to algorithms.

Impact

My proposal will make it very easy and fast for voters to vote on hundreds of proposals based on their interests about the blockchain's development direction even if they have little understanding of the individual proposals. It will therefore be able to engage users that would otherwise shun voting due to the perceived effort or their lack of technical understanding. This will result in a much better mapping of voter's interests onto the proposals than otherwise possible. Taking 5 minutes to vote will make many more people vote than needing to invest hours and hours reading through proposals.

It will also enable users to find their position towards future blockchain development amongst all the other interests that are out there, which will make it easier to engage with like minded people by finding interest groups. This will also foster cooperation among opposing interest groups since they can be seen as groups with degrees of difference as opposed to total antagonism. There will be no "users vs miners", but voters and interests groups who acknowledge each other's position because they are just differences on a scale. This should allow for more fluid cooperation, bigger engagement and more clear leadership and direction of Cardano's development. 

I think this addresses all the guiding questions of the Campaign Brief. To quote the main point:

The community is consciously improving decision-making processes while growing a group of committed and collaborative leaders.

META / Remarks

Thank you for the comments and the many discussions with fellow proposers in this fund! I've amended my proposal accordingly and am happy to announce that already there is communication how my proposal can be tied into other proposals like for example Catalyst Compass ( https://cardano.ideascale.com/a/dtd/Catalyst-Compass/367154-48088 ) and Voter Tool - AIM ( https://cardano.ideascale.com/a/dtd/Voter-Tool-AIM/368984-48088 ) - both of which have nice synergies with my proposal. Voting for them as well as my proposal will foster adoption and further collaboration of all of these.

 A couple of issues that will have to be dealt with in the future: DID and wallets - what is a voter vs. a staking key - how to prevent manipulation of descriptors. Ensuring the quality of descriptors? Will stable, algorithmic clusters emerge? Can useful definitions be assigned to them outside of their mathematical description?

Not all of these issues can be handled in the scope of a single funding round. I expect a team of data scientists, front- and backend developers and community specialists getting involved eventually to drive this to a larger scale.