Polis-style sentiment maps for governance discussions?

Hello,
I was reading mpizenberg 's survey of group-decision platforms and two tools may be relevant to how we do discussion here in the future: Polis and Loomio. (Thinking out loud, would love input)

Right now each governance action gets a thread with comments and up/down reactions, and the real vote happens on-chain (DReps, stake-weighted). The up/down tells us "good comment," not "where does the community actually stand on this proposal."

Polis is the interesting one. It's not a comment tool, it's a sentiment map: people submit short statements, everyone votes agree/disagree/pass, and it clusters folks into opinion groups while surfacing the statements that bridge the camps. So instead of a 60-comment thread you get "here are the two camps, and here's what 80% of both actually agree on."

It would pair nicely with what we already show: the on-chain DRep tally (stake-weighted) next to a community sentiment map (one person, one vote). That gap, "the big stake voted X but the room leans Y," is exactly the thing a governance forum should make visible. Bonus: people who can't post could still weigh in by voting on statements. (On the other hand, I also see this as contrary to the original idea that you can only post on DRepTalk if you hold an on-chain role.)

Open questions:

• Do our proposals draw enough people yet for clustering to mean anything, or is this too early?
• Sentiment as pure signal, or could it feed into something?
• How do we show head-count next to stake-weight without implying the head-count is the "real" result?
• Loomio mostly overlaps with our on-chain voting, so probably not the whole tool. But one small pattern is worth stealing: pick your stance, write one line of why, thread renders grouped by stance. Could be nice for general discussions that have no on-chain action behind them.
• So I'm mostly curious whether the Polis layer is worth building, or whether it needs more discussion volume first to be useful.

References:

• https://github.com/mpizenberg/participatory-budgeting/blob/main/group-decision-platforms-report.md#4-map-of-the-current-open-source-platform-landscape
• https://github.com/katomm/dreptalk.com/issues/109

I'm going to get back to you on this soon, super busy right now. I worked on a version of Loomio, and am very familiar with Polis. Moreover, the social technologies running inside of Loomio (consent-based decision-making) and within Polis (I would describe it as "cross-cluster legibility") are both excellent sources of epistemic hygiene for discussion objects.

Off the top of my head, there are at least 10, and maybe as many as 15 excellent patterns from these two platforms that could be implemented here that would increase the value of the discussions to participants in terms of deliberative capacity. Things like typing discussion contributions, separating clarifications from evaluation, ability to surface a material objection, clarifying elements of disagreement, support for conditional and relational positions, the possibility of exposure ladders for discussion objects, generativity around alternatives, surfacing bridge statements, typing deliberative standing (lived experience, technical expertise, etc) versus voting power, preservation of minority concerns alongside record of the discussion... I'm sure there are a few more.

I'd like to get a recommendation report for you to consider around these things, if you are open to it, and if I can clear my schedule a bit. I'm genuinely appreciative that with all the work on the site already, you are still pursuing genuinely interesting improvements that could allow for plurality.

The biggest challenge, is that a lot of platforms are designed for decision-making (and a lot of the participatory budgeting frameworks are built around establishing the funding agreements themselves, not the discussion around them, so the target there is more like consensus). The nature of DRep voting is not to establish broad agreement, but rather, broad understanding and coherence. A good discussion object shouldn't push a vote one way or another, it should improve the shared understanding even between two voters where one votes yes, but the other votes no. This is the definition of deliberative capacity, a coherent commons is not one in which difference has disappeared. It is one in which the relationships among differences have become clear enough that participants can make responsible decisions without turning every disagreement into an existential conflict.

I thought rather than build a report I should build a static of prototype some of the ideas I was talking about. You can check it out here: https://styg-drep.github.io/crux/

The "Seams" button is basically a list of all the places where it falls a bit short. The idea would be to install it as a module, not as the native script. It speaks for itself in most places, any questions, feel free to ask. It goes beyond the cluster analysis (which I think you have rightly tagged to higher engagement levels) to isolating sticking points (thus...CRUX). It's pretty aspirational in a lot of places, as far as what is actually possible, but maybe there is some inspiration here.