I do find the added complexity distracting and overwhelming.
But that’s less of a problem if such a voting system would only become selectively enabled in contexts where the benefits are worth the cost of extra complexity.
And obviously that feeling of being overwhelmed is only partly due to the increased complexity, and partly because it’s an experimental unpolished feature.
That said, four axes plus extra emojis really is a lot, and I’d imagine a final system would be more like 2-3 axes (or a dropdown with some orthogonal options) than as many as here.
Regarding that point, I do think it’s important that all axes in such a system are reasonably orthogonal. General Upvote, Truth, Aim, Truthseeking, and Clarity all seem far too correlated to be on separate axes.
Also, I’m slightly wary that some of these votes maybe impute too much motive into a comment? If you vote “seeks conflict”, that implies that you think the comment author was intentionally unvirtuous. Whereas if your vote was an Angry Face à la Facebook’s system, that’s obviously at least partly about your own state of mind, and your reaction to the comment. (Not that Facebook’s system would be particularly aligned with the desired LW culture.)
Regarding my impression of this current specific implementation of the voting system, the True-False axis being abbreviated as Truth makes sense, as does the Clear-Muddled->Clarity axis. But the Hits-the-Mark-Misses-the-Point axis being abbreviated as Aim feels incredibly confusing, and same with Seeks-Truth-Seeks-Conflict becoming Seeking.
Regarding the icons, Empathy and Surprise seem reasonably clear to me, whereas I could not identify the meaning of the Skepticism and Enthusiasm icons just from their icons.
Also: The current popup has virtues on the left and vices on the right, whereas the LW voting system is downvote-upvote, not upvote-downvote. So I’d prefer these axes to be flipped in the popup. One could also do something like: <Truth>, <Clarity>, <Truthseeking>, …; and then the “<” is a downvote on the Truth value, and hovering over it gives you a tooltip that explains that this vote means the comment is False. (That said, multiple nested popups and tooltips are a no-go, of course.)
Regarding that point, I do think it’s important that all axes in such a system are reasonably orthogonal. General Upvote, Truth, Aim, Truthseeking, and Clarity all seem far too correlated to be on separate axes.
It took a while to read this, because I’d have said ‘hits the mark/misses the point’ and clear/muddled were the ones that seemed perhaps too similar. Then I noticed that you mentioned those—processing what these words mean quickly is going to take a bit to get down.
A comment, or post, could be clear, and yet, the point might not be. (And I might comment asking there.)
Also: The current popup has virtues on the left and vices on the right, whereas the LW voting system is downvote-upvote, not upvote-downvote. So I’d prefer these axes to be flipped in the popup.
Yes.
I’m fine with the complexity now, but in a big thread, with loads of comments, that are very long...that’s going to be more challenging. Hopefully the new system will make some things more clear, so that the process of understanding gets easier, but at first? It’ll be a lot.
Agree it’ll get better if limited to relevant contexts and polished up.
Agree the axes are difficult to distinguish from one another. True speech, truth-seeking speech, precisely specified speech, and accurately aimed speech are all distinctly important! buuuut they’re strongly correlated so the distinctions are usually only useful to point out on the extreme ends of the quality spectrum, or on very short comments.
There’s an axis? reaction? that is not quite muddled or conflict-seeking or missing the point or false, nor does it warrants skepticism or surprise. It’s just… an ugh field. It’s the category of too much text, too far outside my base context, too ugly, too personally triggering, too why I should even try.
My browser shows does not display the skepticism or enthusiasm icons, I too have great difficulty identifying their meaning.
There’s an axis? reaction? that is not quite muddled or conflict-seeking or missing the point or false, nor does it warrants skepticism or surprise. It’s just… an ugh field. It’s the category of too much text, too far outside my base context, too ugly, too personally triggering, too why I should even try.
Good point. I would not consider all those quite the same axis, but they’re sure orthogonal to the axes we have here.
Here are some potential word pairs to name this axis: Energizing/Inspiring<->Exhausting, Enjoyable/Joyful<->Ugh, Polished<->Mess(y)/Unfocused/Meandering.
Personally, I don’t find these 4 axes to be too much to handle. I don’t necessarily agree that the axes have to be very orthogonal. The point of this system is to promote LW’s desired culture of seeking truth, so it makes sense that the axes are going to have that all in common. The important thing is that each axis should have some significance that is not communicated by any of the other axes- which I feel at least 3 of the 4 axes accomplish (“true” is about whether something is actually true, “clarity” is about how well the thoughts are expressed, regardless of the truth, “seeking” is about demonstrating proper epistemic hygine, (Which overlaps slightly with clarity, but clarity is more about having a line of thought that can be followed, with less emphasis on the quality of the tools of reasoning, while truth-seeking emphasizes using tools that give good results, with less focus on how clear their use is, or the actual resulting thesis).
I’d say I have the hardest time distinguishing “aim” from “truth”, because ultimately something that hits the mark is true, though “misses the point” seems not quite the same as “false”. Actually, now that I think about it, “hits the mark” and “misses the point” don’t really feel complementary to me- ‘hits the mark’ is basically about agreement, while ‘misses the point’ seems to be more about how well the thoughts in a comment understand and relate to the conversation it is a part of.
I would maybe suggest trying to adjust “hits the mark” to also be on this axis- highlighting not just truth, but relating to the broader context of the conversation in a good way.
Feedback:
I do find the added complexity distracting and overwhelming.
But that’s less of a problem if such a voting system would only become selectively enabled in contexts where the benefits are worth the cost of extra complexity.
And obviously that feeling of being overwhelmed is only partly due to the increased complexity, and partly because it’s an experimental unpolished feature.
That said, four axes plus extra emojis really is a lot, and I’d imagine a final system would be more like 2-3 axes (or a dropdown with some orthogonal options) than as many as here.
Regarding that point, I do think it’s important that all axes in such a system are reasonably orthogonal. General Upvote, Truth, Aim, Truthseeking, and Clarity all seem far too correlated to be on separate axes.
Also, I’m slightly wary that some of these votes maybe impute too much motive into a comment? If you vote “seeks conflict”, that implies that you think the comment author was intentionally unvirtuous. Whereas if your vote was an Angry Face à la Facebook’s system, that’s obviously at least partly about your own state of mind, and your reaction to the comment. (Not that Facebook’s system would be particularly aligned with the desired LW culture.)
Regarding my impression of this current specific implementation of the voting system, the True-False axis being abbreviated as Truth makes sense, as does the Clear-Muddled->Clarity axis. But the Hits-the-Mark-Misses-the-Point axis being abbreviated as Aim feels incredibly confusing, and same with Seeks-Truth-Seeks-Conflict becoming Seeking.
Regarding the icons, Empathy and Surprise seem reasonably clear to me, whereas I could not identify the meaning of the Skepticism and Enthusiasm icons just from their icons.
Also: The current popup has virtues on the left and vices on the right, whereas the LW voting system is downvote-upvote, not upvote-downvote. So I’d prefer these axes to be flipped in the popup. One could also do something like: <Truth>, <Clarity>, <Truthseeking>, …; and then the “<” is a downvote on the Truth value, and hovering over it gives you a tooltip that explains that this vote means the comment is False. (That said, multiple nested popups and tooltips are a no-go, of course.)
It took a while to read this, because I’d have said ‘hits the mark/misses the point’ and clear/muddled were the ones that seemed perhaps too similar. Then I noticed that you mentioned those—processing what these words mean quickly is going to take a bit to get down.
A comment, or post, could be clear, and yet, the point might not be. (And I might comment asking there.)
Yes.
I’m fine with the complexity now, but in a big thread, with loads of comments, that are very long...that’s going to be more challenging. Hopefully the new system will make some things more clear, so that the process of understanding gets easier, but at first? It’ll be a lot.
Agree it’s overwhelming.
Agree it’ll get better if limited to relevant contexts and polished up.
Agree the axes are difficult to distinguish from one another. True speech, truth-seeking speech, precisely specified speech, and accurately aimed speech are all distinctly important! buuuut they’re strongly correlated so the distinctions are usually only useful to point out on the extreme ends of the quality spectrum, or on very short comments.
There’s an axis? reaction? that is not quite muddled or conflict-seeking or missing the point or false, nor does it warrants skepticism or surprise. It’s just… an ugh field. It’s the category of too much text, too far outside my base context, too ugly, too personally triggering, too why I should even try.
My browser shows does not display the skepticism or enthusiasm icons, I too have great difficulty identifying their meaning.
Good point. I would not consider all those quite the same axis, but they’re sure orthogonal to the axes we have here.
Here are some potential word pairs to name this axis: Energizing/Inspiring<->Exhausting, Enjoyable/Joyful<->Ugh, Polished<->Mess(y)/Unfocused/Meandering.
if i had to redesign the system right now based on these thoughts, I’d go for 3 sections of feedback.
First, reactions: Skepticism, Enthusiasm, Surprise, Empathy, Ugh, Wrath
Second, upvote/downvote.
Third, rubric breakdown. this is collapsed by default, if you voted Strong in either direction then it automatically opens.
False | True
Muddled | Clear
Irrelevant* | On the Mark
Seeds Discord | Truth Converging
*-possible alternative: out of bounds?
Personally, I don’t find these 4 axes to be too much to handle. I don’t necessarily agree that the axes have to be very orthogonal. The point of this system is to promote LW’s desired culture of seeking truth, so it makes sense that the axes are going to have that all in common. The important thing is that each axis should have some significance that is not communicated by any of the other axes- which I feel at least 3 of the 4 axes accomplish (“true” is about whether something is actually true, “clarity” is about how well the thoughts are expressed, regardless of the truth, “seeking” is about demonstrating proper epistemic hygine, (Which overlaps slightly with clarity, but clarity is more about having a line of thought that can be followed, with less emphasis on the quality of the tools of reasoning, while truth-seeking emphasizes using tools that give good results, with less focus on how clear their use is, or the actual resulting thesis).
I’d say I have the hardest time distinguishing “aim” from “truth”, because ultimately something that hits the mark is true, though “misses the point” seems not quite the same as “false”. Actually, now that I think about it, “hits the mark” and “misses the point” don’t really feel complementary to me- ‘hits the mark’ is basically about agreement, while ‘misses the point’ seems to be more about how well the thoughts in a comment understand and relate to the conversation it is a part of.
I would maybe suggest trying to adjust “hits the mark” to also be on this axis- highlighting not just truth, but relating to the broader context of the conversation in a good way.