FeX: Forum Explorer – Rethinking how we interact with threaded conversations
mcnutt.in
mcnutt.in
Another thing I was doing at the time was visualizing Twitter and Facebook friend graphs, and the first insight there was to deliberately drop the 'ego' node, as it destroys any hope of effectively showing community structure of the friends themselves (it also carries no extra information as everyone is necessarily connected to the ego).
So perhaps a similar thing would be effective for threaded conversations: if you drop the root node, disconnected discussions will become separate trees in the visualization, which will help with efficient use of space, and the tail of singleton comments that no one replied to will become isolated disks that can space-fill the background. Switching to an actual rectangular tree layout as I used in my prototype will also help in that one consistent direction (down) represents nested depth.
Dropping the root node is an interesting idea!
Unrelated, might be fun to surface some kind of iconography to indicate semantics of the content (e.g. flames/smilies/etc)
On sites that heavily rely / encourage getting responses such as Twitter I worry it just encourages posts that are trite, obnoxious and such as those get the most responses.
Granted this is a visualization tool so it's not "responsible" for issues that are largely human / platform issues.
I like that mini-summaries on the left-hand side ("subconversations about XX"), but the graph view seems to prioritise responses with the maximum amount of comments.
E.g. if there is a flame-war going on in one of the subthreads, it will be more visible than a highly rated comment with a few interesting subcomments.
That's very true. This hasn't really been designed around screen readers, making this type of graph + comment tree interface work accessibly well is a really interesting challenge!
>It's impossible to move the mouse pointer from the left pane to the right without changing the visible comment.
If you click anywhere on the graph then it locks the current comment selection. You can then click to unlock!
I think this would be essential. I've been clicking around on it, and it's not entirely clear to me how to work it. I may be an idiot, but I'd really love some instructions.
EDIT: I just now managed to see the "wait I'm lost" link toward the bottom. It took some doing (that's part of the page that gets clipped). I'm now following the walkthrough -- perhaps that will clarify things for me.
Edit edit: OK, I think I have a basic understanding of what's up. It's overly complex for my needs, though -- I think I prefer the normal comment threads.
I'm on it! https://github.com/mcnuttandrew/forum-explorer/issues/67
> Edit edit: OK, I think I have a basic understanding of what's up. It's overly complex for my needs, though -- I think I prefer the normal comment threads.
that's super reasonable! this is an experimental interface, some things are bound to work for some people and not for others
I then wanted to see the context for a specific comment so I clicked on it. Great that worked!
Then I wanted to go back to where I was in the list of comments by the OP. I clicked the back button ... that didn't work.
Ideally either the back button would work (each state has an equivalent URL) or the UI provides a way to return to the last state.
Thanks for the interesting tool!
It might be worth mining the trn user interface for ideas.
The author tried to write a major new version (4) in the late 90s, but from what I could see of the development process (being on a mailing list), the codebase became unmanageable and there were lots of bugs that proved impossible to catch. Usenet was on its last legs by that stage anyway. So that's a cautionary tale: make sure the specification and basic architecture remains clear and clean. (I'm not sure how many of the bugs back then were related to managing ncurses and other fiddly issues that won't arise in the architecture of a web app.)
Threads that the user didn't want to read could be "junked". This could work in the context of HN as a way of sorting posts into interesting and not interesting on a per-user basis.
Rather than just navigating a set of threads, trn gave the user tools to productively keep up with discussions they were interested in, and not have to worry about other discussions at all (after dismissing those discussions once).
Here's a basic explanation of the interface: https://www.einval.com/~steve/software/trn-faq.html
I've got to admit though, I'm a little jaded towards "Chrome only" extensions. Is there a reason you couldn't build this in Firefox?
As web developers, we've really got to stop this practice lest the web become further fragmented.
Anyways, I don't want to come down on you specifically - I realize this is probably just an MVP. You did a great job.
I totally agree, I was just trying to minimize the number of things I was designing for while I was building it.
That said, Firefox addon is in my todo! https://github.com/mcnuttandrew/forum-explorer/issues/45
How does "someone built a Chrome extension" justify you asking them to explain themselves for why they didn't build it the way you want?
OP seems to agree with the sentiment, so I'm not sure why you're bothered.
Questioning the choice (that is, implying it's wrong), asking for whether the reason is "good" as a value judgement, stating outright that "we have to do better", is not perfectly valid. It's rude.
I'm passionate about protecting the open web because I believe a modern and free democratic society literally depends on it. And yes, I know that sounds quite dramatic, but I believe it wholeheartedly.
It's using System WebView. In my case Bromite
Hackernews is written in arc. I also think its interesting the the comment structure is a tree, just like an S-expression:
I really wish my tooling for spell checking while writing jsx-y code was better, i feel like this happens a lot
Is that part of what you're trying to test, or is it just an unimportant artifact of the test page?
Also, I'm not actually on board with the visualization -- it just makes everything more confusing and harder to follow. But that might be my personal idiocy.
JohnFen's may more accurately be this other issue:
https://github.com/mcnuttandrew/forum-explorer/issues/69
I also made an issue for how it looks on Chrome in Android:
https://github.com/mcnuttandrew/forum-explorer/issues/70
On checking Firefox on Android, I wonder if this is the platform JohnFen was trying:
In general, the different layouts might be more to your liking.
But I sure do like graphs.
The circular tree your code generates is very cool. That presentation makes it a lot easier to see how the different threads relate than a flat tree structure.
It's especially fun to see how the topics weave through the conversations. Though commenters mostly stick to one topic on a particular tree, some of the same topics appear on siblings of adjacent trees.
Some of the nodes in a vertical tree, for example here[1], are very close to each other while there is room for them to be a bit more separated.
Comment scrollbar goes beyond window edge.
Firefox 67.0b18 64bit Windows 10.
Overall looks neat and interesting, thanks for sharing.
[0] https://mrandri19.github.io/comments-as-a-graph/?thread=hn3
https://docs.microsoft.com/en-us/visualstudio/modeling/map-d...
I'd like to see sentiment analysis color code the nodes and maybe node to node distance represent response time. And link width represent number of characters in post.
* The circle should plot threads clockwise newest<->oldest (or reversible?)
* Node connection length should be the reply delay in the chain
* Node size should reflect comment size
* Node color/opacity should reflect comment score/vote amount (if present, which HN doesn't have.)