has reached the end-of-life and is now read-only. Please see the EOL announcement for details

#Fediverse does this happen often to you too?

You found a great post, commented to it, only to find out later that a broad discussion was triggered ... but not on the thread branch you commented on.

So if you didn't bump into it coincidentally you'd be totally unaware of that.

I'd like a Watch Post feature that sends me notification of all activity on the thread.

See also:

Would you like to have a Watch feature?

@humanetech I'd like a thread to be *a single thread* instead of the multi-branch mess it is now.

@tinyrabbit I can see why you would want that, & I'm sure I'm not going to say anything you haven't thought of already, but my thinking is this.

Suppose there's already a single thread, not too long, easy to read perhaps 10 or 20 toots. But one of the toots in the middle is something to which I'd like to reply. The existing system allows me to do that, whereas having a single thread would force me to tag my comment on the end, even though I'm replying to something further up.

CC: @humanetech

@tinyrabbit My wish is to be able to reply to multiple toots, thus drawing a sprawling conversation back together. Then, additionally, have the ability to render the discussions as graphs, like this:

I've written a discussion system that lets you do that. It's currently find for small numbers of users, and extremely helpful, but pig-ugly, and very parochial.

CC: @humanetech

@ColinTheMathmo @humanetech To me this sort of discussion tree looks quite horrible 😆 I can't imagine the effort it takes to follow that discussion in all its branches, or how frustrated I would be to see virtually the same discussion taking place along a number of different paths. To me it looks like a great way to *socialize*, but a horribly inefficient way to *discuss*.

@tinyrabbit That discussion tree would benefit from navigation tools, and I have those, and that makes it a real pleasure. Open some branches, close off others, hide nodes that don't really contribute, and you end up with the "Real Content" distilled.

As I say, I don't have the skillz to make those tools more widely available, but the chart, when displayed on a decent size screen, can be scrolled around,and the many threads can easily be followed.

Takes a little practice.

CC: @humanetech

@ColinTheMathmo @tinyrabbit @humanetech

If you want to see a beautiful interface for dynamic discussion in a tree-structured thread, I recommend checking out Euphoria:

· · Web · 1 · 1 · 4

@dynamic I agree that it's a beautiful rendering, but structurally, it simply looks like the usual linear layout of a tree. Is that right?

One of the great things about DiscDAG is the ability to tie threads together. Another is to use space to show the actual DiGraph.

CC: @tinyrabbit @humanetech

@ColinTheMathmo @tinyrabbit @humanetech

Euphoria threading is not shaped like a tree, but it is a tree-structure, communicated by indentation, in the same sense of the directory trees that can be used for file browsing in certain operating systems.

@dynamic Indeed, which means that is suffers the same problem that in a complex discussion you sometimes have to scroll a long way to find a comment's parent.

This is a hard problem, which is why I'm exploring how usable the "actual tree layout" might be.

My experience is that the static version is fine after discussion is ended, and the dynamic version I have is fine provided you use the navigation tools.

Working on it.

CC: @tinyrabbit @humanetech

@ColinTheMathmo @tinyrabbit @humanetech

Tying threads together is not a thing I'd even heard about until you mentioned it in the current discussion thread!

@ColinTheMathmo @tinyrabbit @humanetech

My intuition is that a literal tree-shaped render could be useful for review of a discussion after the fact, but I have a hard time imagining it as an interactive interface.

@dynamic DiscDAG is making progress in that regard. It's intended for dynamic discussions, allowing multiple threads that subsequently come together (where appropriate). It needs better navigation tools, and a better UI in general, but it works well for the appropriate context.

Most examples are private conversations, and those that are public are mostly not good examples, but here are some links.

CC: @tinyrabbit @humanetech

@dynamic Here is an explanation about the Axiom of Choice, and why it's problematic:

Here's a more extensive, free-form discussion about the system itself and some of the problems:

It's almost essential to use Neighbourhood Mode to navigate and follow the discussion a bit at a time, showing that better navigation tools and UI are clearly required.

But I can't imagine trying to follow this in a linear form.

CC: @tinyrabbit @humanetech

@ColinTheMathmo @dynamic @humanetech You have a good point about following branching discussions in a flat thread. In that case compartmentalization and breaking out branches into new discussions when they diverge from the original poster's intention are key. And doing that in social media is... uhm... yeah...

@tinyrabbit A consequence of breaking things out into new discussions is that you can't then tie them together again. If they're genuinely different discussions then that's fine, but if they are different aspects of the same discussion then it's a nightmare. The ability to "fold up" or "abstract out" sections is the thing DiscDAG is missing, and neighbourhood mode is a poor substitute, but when a discussion is large and sprawling, you need to connections.

CC: @dynamic @humanetech

@tinyrabbit I'm looking at being able to select several nodes in DiscDAG and then collapse them under a "Summary Node" which can be expanded or collapsed as appropriate.

CC: @dynamic @humanetech

@tinyrabbit There's a difference between "Shooting the Breeze" (StB) and "Have a Discussion to Reach Conclusions" (HaD2RC). Existing platforms drive everything to StB, and platforms that claim to help with the HaD2RC context just ... don't.

DiscDAG seems to be a step in the right direction. It's needs to go further, but I have none of the skillz, the time to acquire the skillz, or the money to pay people who have the skillz.

So it's stalled.

CC: @dynamic @humanetech

@ColinTheMathmo @tinyrabbit @humanetech

What are your thoughts on Argument Mapping tools? To me they seem much more promising than DAG structured things (I feel like it's pretty unusual to *really* want to pull two separate threads together, and that when it is desirable it could probably be handled pretty well by making a post with links to the parent threads).

@dynamic I haven't seen an "Argument Mapping Tool" for a long time, and I'm sure they've improved, but I've hated every one that I've seen. They have always, to me, seemed to offer nothing over existing linear representations of trees.

And I have frequently use the facility to pull together separate threads. This might be one of those cases where until you have it you don't see the point, and when you've used it, you feel crippled not to have it.

CC: @tinyrabbit @humanetech


Your chart is ready, and can be found here:

Things may have changed since I started compiling that, and some things may have been inaccessible.

The chart will eventually be deleted, so if you'd like to keep it, make sure you download a copy.

@dynamic Here are links to more:

There are many *many* more, but they are private discussions, so you don't have access.

CC: @tinyrabbit @humanetech

@ColinTheMathmo @tinyrabbit @humanetech

The DAG flow charts you've been posting require both left-right and up-down scrolling, which seems like an even larger impediment to comprehensive discussion than linear-format thread trees would be.

@dynamic Large, complex discussions will always need some way to render them, and some way to navigate.

I wish I had the skillz to show you my vision ... clearly I'm just not able to make my point.

CC: @tinyrabbit @humanetech

@dynamic I think I'm just going to end up like Ted Nelson, with a vision of "A Thing" which I know will be better than current stuff, but of which I can convince no one because I have neither the ability to implement it, nor the charisma to inspire others to do so.

CC: @tinyrabbit @humanetech

@dynamic OTOH, maybe it's just crap, and I'm completely deluded. Maybe a linearised tree is the only thing that will ever work, and the best thing that can ever be.

CC: @tinyrabbit @humanetech

@dynamic But try this. Start with this link:

If a node that is currently not highlighted looks interesting, click on the author's name to add it to the selection and see it and its neighbours. When you're done with a highlighted node, click on the author's name to remove it from the selection.

Then start navigating around. You can collapse everything except a single node-and-neighbours by clicking the grey box.

CC: @tinyrabbit @humanetech


I think it is certainly not. But probably you have to target other use cases + audiences more. Here, in microblogging domain, it makes most sense to diagram a complex thread graph after it happened. Dynamic view probably not well-suited.

Specialized audience (e.g. scientists) and use cases (argument maps, decision-making, ...) may be excellent to consider.

It'd also help UX designs if your nodes had more semantics than just <text> e.g. linked data vocab.

@dynamic @tinyrabbit

@humanetech WRT getting an audience, yes, probably these sorts of facilities and capabilities are better suited to a more specialist audience than a generic "social media" context.

WRT semantics and nodes, probably, but I have no front end skills. None. Producing a diagram like this is all I can do, so while what you say is probably true, I can't make it happen.

To be honest, I've pretty much decided to give up, and just use it for myself.

CC: @dynamic @tinyrabbit

@ColinTheMathmo you might describe things in text, no ux designs needed.

For semantics, e.g. the simple fact of a 'node' having a title and subject might give many options for re-arranging the UI, collapse branches, etc.

In special domains more meaningful linked data formats can be used. E.g. in IT context you may label your text with "requirement" and "bug", etc.

Can brainstorm a lot on possiblities, but also talk to audiences on their needs.

@dynamic @tinyrabbit

@humanetech I've tried. I've tried for years. No one seems to get the vision, and this is what I've produced to try to help people "get it", but that's not helping either.

Sorry, I'm currently in a really negative frame of mind about this because I'm spent(*) a lot of time trying to explain, describe, or otherwise communicate, and I've pretty much got nowhere.

If I were a front-end person then I'd just go and implement it. But I'm not.

CC: @dynamic @tinyrabbit

@humanetech Sorry, I know you're trying to help, but I think I just need to walk away from it for a bit ... it's just making me angry and depressed.


My inabilities to get people to see the possibilities is immensely frustrating, and I've wasted hundreds of hours.

CC: @dynamic @tinyrabbit

@humanetech I probably just need sleep. As they say, "Go to bed, it'll look better in the morning" is just another way of saying "Did you turn it off and on again?"

I've heard a theory about why use of Zoom exploded and overtook Skype when the latter was already well established. The theory is that Skype required both parties to be registered. If A wanted to talk to B, and B wasn't registered, then B had to go through the process before they could talk.

But ...

CC: @dynamic @tinyrabbit

@humanetech ... with Zoom person A could be registered, and just send a link to B, and it would work. Much less friction.

So B got convinced, and when they wanted to talk to C, B had the motivation to get registered, then send a link to C. And so it went.

Maybe I just need to find a few people who will use it with me, and then maybe they will use it with others. The problem is this.

I don't want to create a new platform. I want it to leverage off the ...

CC: @dynamic @tinyrabbit

@humanetech ... existing #fediverse protocols and capabilities. But they have semantic limitations that prevent it from happening.

What's more, and I can't emphasise this enough ... I don't have the skillz.

After at least 6 years of trying, the evidence is in ... I also don't have the skill or ability to adequately describe the vision either.

I know ... I'm ranting. It's an overflow of the frustration, and I really am sorry.

CC: @dynamic @tinyrabbit


Things will get better. Don't let yourself get down, and also think of all the stuff you have learned over the years.

@dynamic @tinyrabbit

Show newer

@ColinTheMathmo @humanetech

In another universe, where I would have infinite time to build interactive data visualisations, I would try to implement this thing you're trying to describe with D3.js.

Very curious about your vision though. It seems that you want to visualise a discussion so that you can have a better overview of the discussion and its dynamics, to empower yourself to both better extract valuable information and understand the dynamics that led to the flow of the discussion?

Show newer


That's what I just wanted to advise. I well understand your mood.. have been there myself a couple of times.

You might let go of your immediate goal, but pick up something you find interesting that is still somewhat related to it, like studying radical different UX concept. Master the field. Approach people that think very differently than most others (Bret Victor'like types come to mind)

@dynamic @tinyrabbit

@humanetech Oh, if I had the time. But no, I have full-time work, and this is something I would use in a heartbeat if it existed, in which I've already invested a large amount of "free time", and which everyone keeps saying either "it's awful" or "no, it needs these 6000 other things".

I don't have time. I don't have the ability. I don't have the skills.And I can't persuade anyone.

CC: @dynamic @tinyrabbit

@dynamic It's really, *really* useful in a constructive discussion. Sometimes a single comment to pull threads together is exactly the right thing, and in existing platforms it seems difficult, or even impossible.

DiscDAG (which I suspect you haven't seen) definitely wins for that, but probably won't scale.

CC: @tinyrabbit @humanetech

Sign in to participate in the conversation

the mastodon instance at is retired

see the end-of-life plan for details: