Downvoting ("unliking"), charaterising, and sorting comments

I think this is part of the bigger issue around choices for how to sort comments... right now it's only chronological, but we also have an idea about sorting by likes in the backlog. This would be a lightweight version of Reddit-style upvotes and downvotes.

forwarding a related idea: in phoenix, we generalized actions such as "like", "dislike", "agree with", "disagree with", etc into free-form reactions, which are then aggregated together into "liked x2", "disliked x3", etc which on hover are expanded into names.

^ that's kinda how I imagined tags working

The importance of the 'Like' 'Not Like' or any method is to highlight the discussion comments, intending to reach the best conclusions and then generate the proposals and the voting

I think that 'Not Like" can be very useful in communication, but I also think that it's perilous when people are discussing controversial and/or emotional topics.

Both 'Like' and 'Not Like' are actually quite susceptible to misinterpretation. For instance, I might 'like' something for one reason, but someone else thinks that I 'liked' it for something else. It depends on what they're focused on, and what they care most about. This generally seems okay with positive feedback such as 'Likes', but riskier with negative feedback.

The generic quality of a 'Not Like' can, IMO, really stifle conversation if feelings are already on edge. It's easy, unfortunately, for people to read it as "I don't like you" or "I don't like a single word that you wrote".

It's probably okay with a massive upvote/downvote community 'curating' site like Reddit, but perhaps not so okay with a group which is really trying to stimulate conversation in the support of proposals and decision-making.

We should bear in mind that proposals necessarily do get the equivalent of a 'Not Like' vote. Proposals need that kind of critical attention. In developing or discussing proposals, however, I think it's good to be more specific. For instance, take the time to tag someone and express exactly what you didn't like about a comment they made. I think that taking the time there, and giving that personal attention, often makes communication more efficient overall.

Gred, thanks for your answer

Better explain my reasoning.

In the discussion intervenes any group member is unlimited time and use 'Like' to emphasize or agree with the comment.

Although the days pass, we continue the discussion until it is possible to create a coherent and chord proposal with the content of the discussion.

When a discussion is quality, the proposal normally exceeds the minimum support stipulated by the group to accept as valid the voting

Is it right, is my reasoning about the discussion, proposal, valid voting ?

On the other hand it is true, we can express the 'Not Like' by commenting specifically's identification of the text we do not like.

In publications with several paragraphs long, often we agree with a series of paragraphs, but to disagree with others, really we do not mark anything with 'Like'.

I hope I explained better

Some discussion platforms (Medium?) allow intext highlighting & commenting etc. Allowing a more targeted enrichment of specific content within a paragraph etc. No idea of the tech difficulties of this kind of implementation?

More specific than generic 'like', which usually seems more like a 'wave over the garden fence', a kind of 'Hi there, I hear ya' rather than a direct endorsement.

Also plays a social role. If a respected group member 'likes' etc. others may pay more attention to reading that specific post.

I'm not fond of 'likes' as a discussion/ results-oriented tool, @jgrey , for the reason you stated above. You put it very clearly and directly.

I do accept 'likes' as a basic fact of social networking, bonding, goodwill. So I use them pretty often. I think they do some good. I just don't think they create great focus or efficiency-- especially with long or complicated comments.

I often try to keep my comments short, to facilitate focused and precise responses. Plus, that makes it easier to accurately decide on whether we 'like' something or not! However, some of our topics in Loomio (and elsewhere) seem to require longer comments, to properly express the connected ideas we perceive.

Intext highlighting and commenting is really intriguing. However, if it is pursued deeper than one 'level'-- for instance, to create a whole new sub-thread of discussion-- it can fracture a conversation. On a related note: I think that there's currently a lot of desire in Loomio for people to get the ability to directly "Reply To" comments, but for there to be only one level of "Reply To" comments. That's a concept in Loomio's Ideas Backlog on Trello. (It can use more votes there, if people genuinely approve of it: https://trello.com/c/SAMuMFoc )

So, I'd probably love to see the ability to highlight and comment upon, or to 'like', a specific part of a Loomio comment. I think that could be really helpful in creating conversational focus and efficiency.

@gregorycassel you said ..

Intext highlighting and commenting is really intriguing. However, if it is pursued deeper than one 'level'-- for instance, to create a whole new sub-thread of discussion-- it can fracture a conversation.

I think we have had this discussion in another thread Political use for Loomio and I know it is something myself and @dnephin have been looking at independently for some time. Here is a quote from my post that directly relates to this topic.

If it were possible to split the content of a post, by sentence or paragraph into something that could be “categorized” by the person posting or other group members in some way as to indicate how they feel it relates to the discussion.

supports / adds to / positive …
detracts / takes away from / negative ..
confuses / not helpful / misleading / …
other …
At the moment “liking” a post, especially if it is a long one doesn’t really highlight the parts of the post I feel are most important (to the discussion) or (to me personally).

There is no way to see which points lead to a proposal or decision and which argued against it or lead to other discussion topics being spawned

@carlosgallogaravan Is this the type of thing you were referring to?

I dont have the link, but I remember reading analysis on how down votes (and by implication 'not like') results in lower quality online discussion.

@gregorycassel

However, if it is pursued deeper than one ‘level’– for instance, to create a whole new sub-thread of discussion– it can fracture a conversation.

I think if it's done correctly this fracturing can actually be beneficial, and is really a critical feature for a decision making tool.

Being able to dig into a specific argument or specific point from a proposal is a great way to really vet something, but at the same time you don't want that thread of conversation to overpower the other threads related to that proposal.

If the tool is designed to handle this type of fracturing, I think you get the best of both situations. Participants are able to fully discuss and flush out individual sections, without completely taking over the conversation and ignoring other points. Each thread can continue without any noise from unrelated posts. When some resolution or agreement is reached in a thread, those findings can be used to update the original proposal with clarification, additional findings, or possibly removed entirely.

I think the key to accomplishing this is to abandon the almost ubiquitous "chronological list of posts" approach that has been around since the pre-web era. We should be thinking of discussions as trees or directed graphs, where each response is really a new node on the graph.

There are definitely challenges involved in creating an interface to expose this type of structure in a way that makes sense to humans, but I think that's what makes it an interesting problem.

At the moment “liking” a post, especially if it is a long one doesn’t really highlight the parts of the post I feel are most important (to the discussion) or (to me personally).

I completely agree with this. Like is just not specific enough to provide any benefit to a discussion.

I see the possible list of positions as something like this:

• support - not only do you agree with the point, but you're invested in making it happen. Usually comes along with additional facts to help "support" the proposal.
• accept - not as strong as support, you're ok with it, but are less invested in it's success
• challenge - neither for or against, you think a point needs more background, or more evidence before any position can be taken
• oppose - you're against the proposal
• disregard - you believe this proposal is either misrepresenting itself or is somehow not genuine

Of course this is just one idea. I see a lot of value in a "block" and "abstain" as well. The idea is that the more specific you can be about the position, the easier it is to get a high level view of the state of a proposal or discussion.

Like just doesn't tell you much.

@dnephin , your views on trees or 'fracturing conversations' are certainly valid, but I personally don't think that they're ideal for software which is focused on decision-making. I think that a tangent which takes on any serious life of its own merits a thread of its own, which is easy enough to start at any time. This can keep each Loomio discussion focused upon one proposal at a time, which is the current (very deliberate) design. It also allows Loomio to give people one notification for each discussion, as currently designed, instead of getting 'lumped' notifications for each 'sub-thread' which could develop. (It may be possible to program separate notifications for sub-threads, but I imagine that's not a trivial feature to implement.)

There is more than one way to 'thread' or 'branch'. Loomio allows threading by creating new topics and new proposals, quite easily. I prefer the way it is, but of course, I can only speak for myself.

@gregorycassel you said..

There is more than one way to ‘thread’ or ‘branch’. Loomio allows threading by creating new topics and new proposals, quite easily. I prefer the way it is, but of course, I can only speak for myself.

Can you explain how this current "threading" or "branching" works as you see it. Currently all that happens is a new discussion starts that has no relation, links or navigation to other "connected" discussions unless explicitly linked in the discussion header or an individual comment.

At the very least we need a way to link related discussions and somehow show how they are related, an easy way to navigate between connected conversations and the ability to move a comment to a different "branch" or turn a comment into a new branch if necessary.

@alandavison , I don't see any major problems with linking related discussions in the header/ Context , and/or tagging people who seemed especially interested. Yes, that takes some effort. Yes, it's possible to create more formal structure, but I don't perceive it as a 'need'.

Formal structures, and the connections we personally perceive or create-- for instance, defining some specific 'branch' in relation to a central 'tree trunk'-- are artificial. There is nothing inherently wrong with artifice, nor is there anything inherently right. You apparently perceive a net benefit from added formal structure of one type, while I perceive a net deficit.

Creating 'sub-threads' within a discussion would be a new type of hierarchy. It's a way of saying that "this discussion is part of that discussion". Some people seem to desire that kind of hiearchical organization; I personally don't.

I'm uncomfortable with the idea of moving comments to different 'branches'. That would tamper with the documented flow of an existing discussion. People certainly can link their comments elsewhere, or quote themselves wholly or in any relevant part.

@gregorycassel I understand your position but in essence what you propose already exists, they are called forums. Chronological lists of posts on a particular subject or topic and yes many do have "voting" and even "survey" components.

I am using / supporting Loomio and providing feedback because the traditional Forum is not well suited to collaborative decision making. How does your view of Loomio differ then from a standard forum?

I think Loomio already differs fundamentally from a standard forum, @alandavison . The biggest difference is proposals, based on consensus theory. Here, we've simply been talking.

@gregorycassel the proposal is technically just a poll attached to a discussion. Ok it has a longer description than most forum polls and the ability to attach comments to your "position" but there are also forum plugins that do this.

@alandavison , I think Loomio was designed to be simple and accessible, and to be very much focused on the proposal feature, and I agree with that. I also agree with everything I wrote here earlier regarding topic/discussion focus. However, I don't work for Loomio, so it may be better for you to get feedback from team members.

As one more note for now, @alandavison , I think we should all remember that Loomio is open source software, and people will be able to modify it externally, regardless of what the core team chooses to develop or offer. This is not meant to dismiss your concerns, but to simply give them some context, in relation to whatever perspective the developmental team has on this subject.

@gregorycassel - I understand that Loomio was designed to be simple and accessible. This is the main reason I am here using it, supported the crowdfunding and providing feedback on how to extend or improve the platform.

Imagine being a new user coming across the "Political Use for Loomio" topic if it had or does continue to grow as it did in the weeks leading up to Christmas. It was running over 5 pages in less than 2 weeks. How would that look after 10 weeks or even 52 if you attracted experts from around the world into the discussion.

Do you think anybody would invest the time in reading all those posts, then manually copying and pasting links, following links to other discussions (that then may lead to more discussions) simply to enter the debate.

Personally I would like to see a tool or platform that allows me to view a categorized list of "discussions". I would like to know if that discussion is still active, how many posts had been created since I last visited, notifications of specific references to my posts but most importantly who agrees with or disagrees with the proposal and why.

I would like to separately read all the posts or points for or supporting a particular discussion. Then all the points against.

Using the terminology suggested by @dnephin I could then tag which points I support (am invested in making happen), those I agree with or accept, challenge, oppose or think are unrelated to the discussion or misleading.

I don't have a problem with chronological posts, what I think most of us who want more from a collaboration tool are suggesting is additional meta-data that can be used to filter the posts into "streams" that offer different views. This does not have to over complicate the interface and certain things that would be nice to have like a per-paragraph highlighting and annotating are not possible using a mobile web interface.

"Do you think anybody would invest the time in reading all those posts, then manually copying and pasting links, following links to other discussions (that then may lead to more discussions) simply to enter the debate."

-- No, @alandavison : I would start over and try something fresh, which would not require any research at all for newcomers. I think that that creates accessibility and inclusivity.

@gregorycassel so how would you avoid simply repeating or restarting a debate that may already be well on the way to solving the problem you are simply going to start afresh. Is your view of Loomio limited to small scale, small group interaction that can be posted, debated and decided quickly?

No @alandavison , my view of Loomio is not limited to small scale, small group interactions that can be posted, debated and decided quickly. My view of collective learning is quite different from that.

I don't have time to engage this issue extensively now, but Loomio already allows specialized discussion through subgroups and nested subgroups, as well as tagging people and cross-linking threads. There are many ways to use such tools.

As I mentioned earlier, Loomio is open source, and people can develop many tools to extend its functionality in specialized ways. As I also mentioned earlier, it may be better for you to get feedback from Loomio team members. I probably won't be able to reply here for quite a while now.

Relevant to the original topic of this post:
How The “Down-Vote” Leads To A Vicious Circle Of Negative Feedback

@alanna Very with you on that article and the viewpoint it promotes. Online communities can go negative so quickly-- infrastructurally speaking we have to do all we can to shape the discourse in a positive manner.

@gregorycassel @alandavison Re mapping the visual grammar of divergent online discussions. Just wanted to bring the following site to your attention.

It seems to do a pretty good job at melding mindmap, content management, and conversational divergence. Behaves almost like a visual tiddlywiki. (One of my favourite tools for many years)

Debategraph, is an Australian software project, seems free & open, designed for political discussions etc (co-founder, Peter Baldwin, is past Aus Federal MP & Cabinet Minister).

Do persist in exploring past the first page. The onboarding experience isn't that great. But if you start to play with the options, especially views, it has very diverse & comprehensive functionality. (Box views, show all works best for me)

Break out the sidebar into page view to better see content. Has lots of great functionalities but you have to dig and play a bit to discover them.

Again, not a great initial UX. Action links/buttons are often almost hidden in smallest text & obscure locations. Persist in playful exploration. There is a lot hidden in the fine details. (eg. Message flows are under the 'Stream' tab)

It's been around for a few years so a bit surprised it is not more well known. Seems to be still alive and in use but not overly active. Haven't quite worked out its 'business model' yet? Seems very user friendly. For example you can replace the site branding with your own, complete with URL redirection to your home page etc.

Site probably needs a little updating. Looks a little 'grandmotherly' with its pastel colour schemes etc. Morphing it with the proposal driven capability of Loomio would be very interesting.

Debategraph

Try looking at an active discussion to get a better feel.

You might like

An Open Source Pattern Language (re)generative of Commons

Or

Civilization Project

@simontegg is working on a project around argument mapping. Seems relevant here?

@Gray @alanna @simontegg - Helene Finidori of the first DebateGraph that Gray cited might be interested in discussing how Loomio might evolve into argument mapping.

We're thinking of something sorta related to argument mapping for early-stage research and development where a group of people are trying to evaluate alternative paths for a project, and would be very interested in ideas along these lines.

I'm afraid I can't look closely at Debategraph now @Gray -- much time pressure. The general design there, and the versatile view options, look strong. I'd have to study the tagging and search features to get a good feel for what it's ultimately capable of.

@bobhaugen , @alanna , @simontegg , since you've all been involved and/or tagged here, I want to share some general views on the subject of characterizing and sorting conversational contents/arguments. I'm not sure how relevant these concerns are to anyone's specific views on content management and argument mapping.

I'm leery of conversational divergence/fragmentation if and when people use it to formalize or 'mechanize' our deliberation and decision processes. For instance, people sometimes seem inclined to construct macro-agreements as a supposedly 'logical' or necessary consequence of micro-agreements. I don't favor that approach to group process. I like to approach whole issues holistically, and not as a result of any syllogism or artificial, complicated structures. Complication has its place, yes-- in processes which must, or ought to, occur sequentially.

Written 'natural language' statements are limited, but they're more flexible than any fully formalized structure. Artificial structures often strip the nuances and intensities from our feelings on isolated 'parts', and from our feelings on how those parts are related. I don't think of life or of nature in any mechanical or potentially mechanical way.

A micro-agreement is perfect for a micro-action. It may or may not be relevant to a genuinely consensual macro-agreement. Micro-agreements can lead to the problem of design by committee . Formal structures involving hierarchies, nodes etcetera are useful (and practically necessary) for reference and research purposes, but I advocate minimalizing them with respect to group identity, agreements and goals. I'm pretty radical that way.

This all relates to my pro-Loomio thoughts on collaboratively generating and handling yes/no propositions. I don't think Loomio can handle all group needs. Alanna , I hope you don't mind me drawing a connection here with your MiX article on collaborative funding -- it catalyzed deep change the way that I think, and not just about funding. I think of Cobudget as a real game-changer in terms of … well, gamifying a group's proportions and intensities of interests. It creates organic process. That organic process does, in some cases, relate to 'make-or-break' goals, which can theoretically be reduced to yes/no propositions. However, that's far from always the case.

To try to summarize a bit: I think people often want to create more collective formal structure than we genuinely have, or ought to have.