75
votes
Comment vote counts are now visible again
It's been just over a week since the experiment with hiding comment vote counts was started, so as promised I've turned it back off now, and votes are visible again.
There was already a fair amount of feedback posted in the original topics as well as random other threads over the week, but this can be the main place for "postmortem" feedback now.
I won't bias the discussion in a particular direction by giving any of my own thoughts just yet, so let me know what you thought of it and whether you think we should consider making any similar changes more permanently.
My main takeaway from this experiment was that Tildes felt smaller and less populated to me--more lonely and vacant. This wasn't based on post or comment frequency (I'm pretty sure my personal usage actually increased this week) but was based on the fact that I had no indicator that my or anyone else's posts were being seen. I went from knowing that a particular comment was definitively read by, say, 15 people, to wondering if anyone had seen it at all. Furthermore, while I continued to vote on things as I had before, it saddened me a little to know those would go unseen by those people. I didn't realize until this week how often I was using voting to be a digital thumbs up or pat on the back to other users.
Removing this means that the only way to engage with other users is to reply, and I think it's impractical to assume everyone who wants to reply will be able to do so. I know that, on any given day, I'm only able to comment on a handful of the things I want to because my time is limited (and because I'm wordy to a fault). In the absence of a meaningful response, voting is still a way of letting other people know that I'm reading what they're writing and I think, whether or not I agree with it, that it's still a valuable contribution to the conversation. It's not primarily about validation but about making the audience visible to others in the first place.
As such, I think the biggest argument in favor of including and publicizing voting is that it is a time-efficient method of participation that makes our community visible to its participants. Without it, we run the risk of our community being defined not by what most of us collectively deem valuable but by those who comment most frequently, as they become the most seen. Without it, the only way to be seen is to use your words, and while there are definitely some benefits to that, I don't think it scales to the kind of site we want Tildes to be. Voting is a counterbalance to commenting in this way, allowing people to still have a say and be present in conversations without having to overcome a high barrier to entry and spend limited resources in order to do so.
Now, with all this said, given that I'm in favor of it, how can we "fix" voting (if at all)? I have two ideas, plus one bonus idea. None are fully solidified and all should be treated as loose frameworks from which to build--or tear down outright. I've been in my head about this all week and can't see past the limits of my own beliefs. As such, I could use some outside perspective, and implore people to tweak/critique these heavily.
Idea 1: Positive Labels
This is a possibly overthought one. It's complex, and would require a lot of shaping and finesse, but the basic premise is the addition of a robust set of positive labels.
These would be a list of comment classifications we value: stuff like
interesting,well-written,poignant, and so on. Importantly, it should have something to signify posts you disagree with but that are still valuable to the conversation at large--things likechallengingorthought-provokingor maybe even something likechanged my mind.The idea here is that a user can label a comment using any combination of positive descriptors from the list, and doing so acts as a vote for that comment. This, also, would be the only way to vote for comments, so we would no longer have the simple ambiguous
Votebutton.Up to this point, it's similar to what we have, but I think what could make this work is that publicly, we should only show the aggregate score for a comment and not its labels. Instead, the labels themselves could come into play via sorting. So, if I go into a thread, I can sort by
poignantand the posts with the highest score for that label (not their aggregate score) would be at the top. Same goes for any other positive label.If implemented correctly, unpopular opinions that are well-stated or thought-provoking would still receive high aggregate scores because they would still get labeled and therefore voted on. It would also reduce the idea that votes = agreement, as we are still encouraged to label things we disagree with provided they meet a certain standard. Furthermore, it would level out comment hierarchies a bit because it would eliminate the idea that there is a single "top" comment, as there could be a number of different ones in the top spot depending on your sorting label of choice.
Within this method I think transparency at the individual level would be good, so that each user can see the breakdown of labels for their comments and interpret specifically how the community responded, but I do not think that should be publicly available, as it would create even more axes by which to compare comment scores.
Idea 2: Rebranding Voting
After spending too much time overthinking the previous method, I decided to challenge myself to go the simplistic route and see if I could identify the easiest possible overhaul with the maximum possible benefit.
My simple solution is to rebrand "voted" as "contributed." People have talked at length about voting as representative of a whole mishmash of different motivations, to the point that it becomes an imperfect heuristic of everything it tries to encompass. Is it a measure of agreement? A tally of read receipts? A log of likes? The answer is simultaneously yes and no to all of these.
Instead of attempting to distill these down or push a particular usage, let's lean into the ambiguity and reduce down voting's role to a simple binary distinction: a comment either contributes to the conversation or it doesn't. It's up to each individual user to determine what they consider "contributing." For some, that'll be posts that reflect their viewpoint. For some it'll be posts that make them think. For some it'll be any post that isn't malicious. Whatever the reasoning, seeing
Contributed (22)after a post simply means that 22 people felt like it added to the conversation. Nothing more and nothing less.I realize it's not much of a change from what we have now (which was what I was going for), and I think the primary change isn't anything structural but instead a cultural shift to how we see voting that the language change can only somewhat influence, but I think it would formally codify that a vote means nothing more than what it is: an individual assessment that a comment contributed to the conversation.
Idea 3ish: Vote Ceilings
I have one final idea that is method-agnostic and applies only to how we publicly display votes.
Knowing that social comparison is a real and often detrimental phenomenon, what if we had a very low public ceiling for votes? As in, votes are displayed only up to, say, 10 and then they stop there. This helps guard against the implicit power dynamic and ambiguity when you walk into a thread and see a
Voted (62)comment followed by aVoted (11)one. Is the 11 a worse comment? Or was it simply posted later and therefore had less readers? Or is it a more unpopular opinion that was nonetheless well-worded enough to garner 11 votes? The current method allows for all of these narratives.If, however, we make 10 votes the ceiling, both would display as 10, and we would know that there was at least sufficient support for both to get them up to that level. By keeping the ceiling low, many comments would effectively max out and have the same score, but I think this could actually be beneficial.
It could be paired with sorting based on ceiling score rather than aggregate score, with ties being randomized. This would break up the singular focus on a singular top comment and more evenly distribute readers across different threads, leveling off discussions significantly. For example, in a thread with top-level comments at
43,22,14, and11, a ceiling of 10 and random sorting would allow each of those to be in the top spot.Keep in mind that I chose 10 simply as a random example and am not committed to it in the slightest, other than I think that a ceiling should be fairly low. I do think that if we consider this it would require a more robust discussion about what the precise ceiling should be. The idea isn't to just level everything such that all comments reach the maximum but instead that the ceiling should allow for late, great comments to still rise up and gain similar real estate to early, good comments, as well as cut off vote-based comparisons between most comments.
I (and n others :p) vote for the 3rd option. All I care about is whether many or few people think a comment is valuable to the discussion. I don't need to know how many exactly. Just that no one, a few, or many people do.
This could be displayed as either the numerical or word values for:
Why not make it dependent on the overall thread activity then?
Let's say we have a random topic. Let's say the top comment has 100 votes. It would then be used to compare the rest of the comments. So a comment with no votes would not show anything, a comment with less than 10 would show "a few members voted for this", less than 50 "some members" and more than 50 "most".
It's just a suggestion, wording and specifics can be changed, but by anchoring the ceiling on the local top comment you avoid problems when the site grows plus it can be recomputed everytime a new comment is posted.
Does vote ceilings have advantages compared to hiding votes for some time?
Because I prefer to see the actual number so that it can be compared between different threads.
I agree with your first paragraphs, but none of your ideas really interested me. Vote ceilings seem very arbitrary, especially as tildes grows and it would become easy to hit small ceilings, essentially removing the purpose of voting.
I personally don't think voting is a problem here on tildes, but I'm open to having my mind changed if someone can put together a good argument against it. Voting, and being able to sort threads by votes, is the big improvement sites like reddit made over old sequential forum styles. Have you ever been to a large forum that gathers huge posts per topic? Imagine coming back to the site in the evening and an interesting topic has over 100 pages of comments with 25 comments per page. Ain't nobody got time to read all that.
The first option is interesting and I like it, though I think the fact that it is a step or two more complex risks unforseen behaviors. It definitely warrants more thought and discussion.
The second idea I think is great just because it is in fact simple, and I think it would be a worthy experiment for @deimos to try for a week like he did this one.
The third option is also interesting. I had had some similar thoughts along the lines of a way of displaying that a post was interesting or worthy, and not make the scores competitive. An example that comes to mind is Overwatch's social scoring system where you can only see the percentages of the different accolades given to a player. Which might combine well with your first idea.
I like this idea.
This is a great quick & easy improvement.
I also like idea 1 and I don’t see any disadvantages. It can be combined with ideas 2 and 3 and adds additional functionality.
I like the first idea, but I think it would take a lot of thoughtfulness to develop a good one. It dovetails nicely with json’s suggestion as well. Rather than showing votes, you could let comments “level up” by collecting certain endorsements.
I’m not as much of a fan of #2 though. Needing an affirmative motion to decide whether there was a contribution or not risks biasing things. It’s the yelp and Glassdoor problem where the people with bad things to say are more likely to act on it than the ones who have no strong feelings either way. It winds up being a fair bit of input for every comment you read to use it as intended, and I don’t think most people will.
A “read” marker might work the way Instagram does it. It can show you views independent of reactions for videos. It may not work with a text post since they take more active work to absorb than a photo and you can’t track completion like you can with a video. But maybe you can do it for posted links?
I think the vote ceiling in option 3 should change depending on the average amount of votes (or just something dependent on the amount, doesn't have to be the average)
Eg on a post with 5 comments. With, 1, 7, 5, 10 and 11 votes each. The ceiling may be set to 7 votes.
Then later on, as these comments gain votes over time and new ones are added, they may have 2, 3, 7, 10, 8, 21, 15, 30 and 32 votes. This would set a vote ceiling of say 14 votes.
I liked not seeing vote counts. It prevented them from influencing my opinion on a comment, and made highly voted comments that I disagree with less likely to make me want to angrily comment a rebuttal.
If this were to become a permanent change, I think it would be useful to be able to see how many votes your own comments get when viewing your user page. This might give users that positive feedback we crave so much without the negative aspects of public vote counts.
Even with vote counts hidden, comments should be sorted like normal so popular comments still rise to the top. I'm guessing this was still the case during the experiment, but I want to bring it up because I still think it is an important aspect of how sites like this function.
I also really liked not being able to see votes. Personally, I'd prefer if votes were hidden to everyone other than the creator by default for a certain period of time, perhaps a week, so that you could eventually receive feedback on your comments, but wouldn't have to deal with the social pressures associated with a popularity dynamic in internet comments while actively engaging with the community. I felt like my comments were less composed and more naturally spoken when I wasn't thinking about that.
I'm a new user who wasn't here last week, but to me this seems like the best solution. A similar change is being made on YouTube with subscriber counts - removing the "instant gratification" users get with an action by truncating numbers to 42K rather than 42,351 for example.
Here, if people are voting for a comment, are they voting simply because they agree or because it's a well written, high quality comment? With this figure hidden I think people will be more convinced to speak their mind instead of say what will get them the most votes, and on a discussion site I think that is quite important.
Agreed completely. It was very refreshing not being influenced by other people's votes on other people's posts. And IIRC there was someone who made a good point about arguments in comments being split between "winning" and "losing" by vote counts, hiding votes solves that problem completely.
But at the same time I missed not being able to see my own votes after the first day or so. I would not oppose bringing it back with the one caveat you said, keeping vote counts visible through the user page alone.
I like this middle ground. Seeing votes here hasn't been an issue for me, but recently I had a problem with it over on Reddit.
There was a world building discussion, and I posted a snippet of lore that I had been working on. A few hours later I went to check the thread only to find that I had no upvotes or replies. Meanwhile, a number of other people that did the same after me were upvoted.
It's hard to not take that as some kind of personal insult. Soon after I caught myself inwardly being a dick to others in the thread. I was reading through others' stuff and thinking "this isn't any good. Why were they upvoted and not me?"
Without the actual numbers I would just see that no one had replied. It might still have been discouraging, but I don't think that knee jerk negativity would have been there.
That sounds like you were either shadowbanned, or automod whacked your post for some reason (like tripping a language filter). You'd have to log out of reddit and view the thread that way to see if your comment appears or not.
I've had automod delete posts before, so I did check that. Turns out the problem was just me being new to the sub. Conversation there was much slower than I expected. After a few days I had a couple of upvotes.
Honestly, I really want to see disagreements here. It's good to see ideas from multiple angles. Of course, that's assuming that you're writing well-reasoned responses and not immediately giving into your anger.
It's not about avoiding disagreements, it's about not putting the users brain in a strong emotional state. Seeing a comment you strongly disagree with have a ton of upvotes can make users respond with a lot more salt, rather than disagreeing in a more calm and collected manner that is more conducive to productive conversation.
I agree with you, I preferred the week without. There's much to say - and I said a lot already in the previous thread - but between the two options we've tried I definitely preferred not seeing them.
edit: Less than 24 hours later and I'm already back to chronically checking my userpage. Oof. :(
I missed vote counts mostly because I couldn't get a read on whether my comments were any good or not. I use that feedback to calibrate myself.
I...don't do that. I post what I need to say rather than what I think others would want me to say. On reddit I do this though if I get a negative comment vote because that should signify that my comment was off topic or unwanted within the discussion.
When you're in person, you use all sorts of verbal and nonverbal cues to determine if the conversation is going well or going sideways... that's more what I'm getting at here. I'm not trying to conform so much as make sure I'm being a benefit to the conversation.
90% of the time I’m downvoted on Reddit it’s because I expressed an opinion that’s contrary to the political slant of the sub I’m in, like daring to say something nice about any democrat other than Bernie in leftist subs or pointing out that Econ 101 isn’t the be-all-end-all in of policy in the neoliberal sub.
Most of the time I’m downvoted on Hacker News I don’t even know what the deal is. It seems completely random.
Downvotes in most platforms are less than worthless at providing decent feedback about whether you’re contributing productively.
i think @rkcr's point is more that they use it not specifically as feedback on their opinions so much as how to express those opinions to other people. i sorta expressed a similar sentiment in a post that's buried in one of the threads around here and which i have the text to but can't be bothered to find the link for:
I'll more write about this later, but this purity trend that's so popular on Tildes bothers me.
I don't want to be influenced, I wanna read the threads with fresh eyes
Why votes at all?
Maybe we should randomize sorting!
Groups are ghettos, let's randomize them too!
If you think about it text itself is heavily biased! We tend to pay more attention to starts and ends, and that is not fair. Let's randomize the sentences!
The mere fact that language is sequential is an attempt against our liberty. I wanna read with fresh eyes, so let's randomize reading direction too!
That's of course a weak attempt at
reductio ad absurdum, but NO. If I didn't want to be influenced by other people I'd live alone on a mountain without any books or devices. If I didn't want to be influenced, I would not have accepted an invitation to a forum full of people. I care about feedback. EVERYONE cares about feedback. That's part of being human and there's nothing wrong with it. I care about votes the same way I care about what a physical audience thinks about my statements. Actual audiences manifest their agreement or disagreement via body language. On the internet we use votes. If the mechanics are good and the system is fair, what's the big deal?That's too big of a leap. I want collaborative filtering from the community. I want to benefit from community signals about what is interesting and what is noise.
But a vote means different things to different people. Sometimes it's "I agree with your opinion", or "I want to promote my ideology", or "Let's push this thread higher so my reply shows up above the fold".
Once in a while, a vote means "I disagree with your opinion, but that's a well reasoned argument that taught me something new."
Munging all those conflicting and contradictory sentiments into a single number is pretty meaningless.
And when you realize that vote count is driven more by total views than actual quality, it's doubly meaningless. If one post has 20 votes in 1000 views, while another post has 15 votes in 100 views, which one is better?
Ultimately, I want labels to capture what's good or bad about a post, not just one number. And I want to choose which users' curations influence my feed, and which users' curations I want to ignore. That kind of optional personalization might fix the flaws of one-size-fits-all voting.
Perhaps we could abstract the vote score. The labels can be expanded and replace the vote button, for example, with labels that more accurately assess why a post is being voted on. We know if a label is positive, so we can use that to tally a score of all the labels that count positive as upvotes. It's taking the temperature of multiple axes of reasons but still giving a vote score.
This gives us the power to develop systems that key into the meanings and functions of the labels rather than into one static vote score, which is now just one byproduct of the label system. Over time we're likely to see other ways to use this for other purposes. People want to vote honestly, in the moment they'll choose the label that corresponds to how they feel about the post. They'll suggest labels over time that they feel are missing, and I expect someday groups might even have their own specific labels local to those communities.
It does kinda get the vote out of the way, and the thinking about voting (which we all seem to do for many different reasons) out of the way. It also means you don't have to vote and label as two separate activities - now they are the same activity, which should mean more people using the label system. We also keep the feedback effect going that many users were missing, since the score is still there.
Well, I was using rhetoric. I was trying to stress a point.
There is no mechanic or social interaction, digital or not, that is able to correctly convey all the sentiments involved in the exchange of ideas. They're necessarily incomplete and prone to errors, but useful nevertheless.
This is not about statistically sound feedback, but having a sense of community that comes from the subjective feeling that what you wrote had an impact on another human being. I'm open to alternative mechanics that are able to provide that, but right now I don't get what's so bad about voting.
I've noticed that rhetoric and other non-literal language doesn't
flywork very well here on Tildes. Any time I try to get poetic or literary, someone will take my metaphor literally, and I end up having to defend something I didn't actually intend - like you did here. It's an interesting phenomenon. I almost never encountered this behaviour on Reddit, but it happens a lot here on Tildes.Yeah. Ironically, Aristotle laid the foundations for both logic and rhetoric, which were used in conjunction in the art of public speaking. But we had to screw everything up by separating those disciplines, so now programmers distrust persuasive language and poets can't do math :P
Early pre-subreddits reddit was more like that too. It's just one of those things that happens when an ocean of STEM crashes into a new website.
I wasn't going to share my theories about why non-literal language doesn't work well here. I figured I'd get lynched if I did!
However, I believe that another contributing factor is that there seems to be a high proportion of people on Tildes who use English as a second language, so it's understandable that those non-native speakers might not recognise metaphors and allusions and slang and common sayings.
As a programmer with a degree in film and an interest in philosophy, I feel like an universal translator.
i'm just going to repeat one of the points i made in the big thread in this vein that concurs with this:
i maintain that a lot of people on here seem to be incredibly reactionary toward certain ideas which have gone poorly elsewhere, without seemingly considering whether or not it's the implementation--and not the idea--that's actually the problem. i honestly don't see voting as an issue unto itself, i see shitty implementations of it as the problem on most social media sites that do it.
on reddit, for example, votes are mostly problematic because they're trackable, which has made the culture around them toxic; because they have material impacts on certain things like where certain accounts can comment in part because of the former and also as ways of community control due to other gaping issues with reddit; and they've as a whole essentially been transformed into a form of social capital by the community instead of purely being a system of community regulation and curation. and yes, those are valid problems--but they don't exactly indict voting as a system, they indict reddit for poorly implementing a voting system. tildes removed the downvote and tracking of votes, and nearly all of those problems immediately ceased to be applicable to the voting system here.
Yes. Stressing what you said, I'd say that these simple differences make a huge difference on Tildes:
I don’t think cumulative karma and social capital is what makes voting toxic on reddit. It definitely encourages karma whoring, but the toxicity comes from people using votes as a way to silence people and the people who perceive themselves as being silenced reacting negatively. It’s like a silent way of shouting someone down when they get downvoted, and once the norm of only downvoting for genuinely non-contributing posts goes away that’s what you’re left with.
I think it’s a natural problem that comes with scale when you have upvote/downvote systems. When Eternal September kicks in you can’t count on norms or mores to keep people behaving productively or refrain from gaming the system.
Yeah, it is reductio ad absurdum. Tildes is still in alpha. If the site isn't trying things that make you uncomfortable, I don't see how we're going to end up any different than just another clone of Reddit.
I don't think that is true at all. Mostly because you're talking out of assumptions that don't correspond with my expectations for Tildes. I don't have any wish tor Tildes to be like Reddit. In fact, I think that the way Tildes deals with sorting, votes (other than this test-week), tags,
exemplarytags, etc is vastly superior to Reddit.There's no guarantee that making me uncomfortable would make Tildes better. Maybe I have great ideas and desires, so making me more comfortable would actually make Tildes much better. But you don't know, you can't read my mind ;)
Yeah I thought I was going to get into how this impacts the community from a broader perspective but since that seemed basically identical to before (if a little worse but I don't honestly believe that's the reason) I'll leave that to the side and just put some personal thoughts:
Site is lonelier than ever without the passive feedback you get from votes.
Time spent checking feedback is now replaced with guessing at what the votes really are.
Threads are far less readable, this gets to be worse as a thread grows.
I care about what people here think (even collectively) and it sucks not having clear indications of that all the time.
Labels make everything confusing and introduce even more doubt into why a comment is where it is. There also seemed to be more superfluous exemplaries but I'm willing to concede that that might just be me seeing it that way.
While I'm far from a lurker, I always read a lot more threads than I comment in and that experience feels much less rewarding. It feels like I'm just reading through a page, not engaging with it as part of the community. Clicking a button to express my thoughts or whose contribution I find valuable makes lurking feel better and I don't think de-emphasizing lurkers in a thread is a good idea.
I don't understand this. I've seen a few people say it's hard to read threads without the vote counts being visible, but I don't get it. With or without votes visible, you see the same comments, with the same text, in the same order. The readability should be exactly the same with or without the vote counts. How is it harder to read a thread without the vote counts?
This probably isn't a satisfying answer since I'm mostly repeating what has been mentioned elsewhere, but the contextual information of how a comment is being received relative to others helps me understand where I am in a thread. Similarly, they can act as landmarks, like if there's a big dropoff in votes in the middle of a thread then there's meaning attached to if a comment is above or below that dropoff that I miss out on. Votes are also easy to scan and remember...if I close a thread and come back shortly after to find a comment, it's a lot easier to find it if I know both generally where it is located and the number attached than if I just have a general idea of the location.
Another part would be that "reading a thread" to me includes how efficiently I can extract the stuff I'll remember from it, and part of that is "what does Tildes think about this topic?" regardless of if we can see votes or not. Not being able to see votes just makes that process longer and fuzzier.
(I imagine some others also mean the redundancy within threads which makes them messier was made worse by people feeling a greater urge to comment and repeat what others think. That wasn't my point, though.)
Most of what you describe is about contextualising comments, rather than merely reading them. But thanks for explaining. At least you've shown me that what I think of as "reading comments" is not what other people are referring to when they say "it's harder to read comments". They're not talking about whether the text on the screen is readable. They're talking about it being harder for them to contextualise and interpret what they're reading.
Why do you say labels make things confusing?
Because they can radically change the sorting of a thread. This isn't a big deal when you can see that a comment with much less votes is high up because of an exemplary, but when you don't know the votes...is it up there because it got boosted? Is it so overwhelmingly popular that it's that high and exemplaried? Suddenly there's doubt even added to the most clearcut top of the thread popular comments, since you're now guessing at votes.
@Deimos myself and others felt like we voted less when votes were hidden. Can you run the numbers to see if that's true and to what degree?
It's hard to compare meaningfully overall, since it depends on so many variables (number of comments posted, how popular certain threads are that day, etc.). Some days had more votes than the week before, some days had fewer.
Here's the average number of votes per comment (total number of votes divided by total number of comments) for the last two weeks, and some of the highest days were from during the experiment (which ran from late June 6 until late June 13):
Let me learn me some plotting.... here is the result, below is the code. It looks pretty random to me, but it is kinda interesting that the two peaks (one slight, the other not so) coincide with the two Sundays.
Yeah, seems pretty noisy. Thanks anyway!
Median isn't very exciting either, they're all just 2 or 3.
4/7 of the days during the experiment had a median of 3 and only 1/7 of the days from the week before did, but looking further back at earlier weeks it doesn't seem uncommon to have 4/7 be 3.
my observations: i stand by everything i said in the first thread about votes, and honestly i've seen absolutely nothing in this endeavor which would sway me toward not having votes being desirable over having them, because nothing to me has changed at all through the course of this experiment except that everything was more inconvenient, threads became harder to browse with size than they are normally, and the typical feedback mechanisms which voting encourages that i touched on in my comment were stifled. i think the problems people associate with upvotes remain overwhelmingly with the community rather than their technological implementation, and i think this all has mostly just reaffirmed to me that the notion that no upvotes is better is people being reactionary. i genuinely did not observe any truly worthwhile changes in behavior between having them and not having them for what was temporarily lost.
also, as another observation just now: the announcement thread is a pretty good illustration of one problem with not having votes, which is the near complete inability to tell how popular an opinion is.
without votes, you can sorta kinda ballpark things based on sort, but if you went into the announcement thread without votes as we all did for 7 days, you'd probably assume it's a pretty even split of opinions in the community based purely on the parent comments and their replies, since there is indeed a very even split in the vocal commenting base in that thread. you'd also probably assume then, based on that, that the votes underneath will ultimately reflect a split of that sort.
however, now that we can see votes and get raw numbers on how many people "like" an idea in an abstract sense enough to vote it up, it seems quite likely that community does skew a distinct way on votes, because my line of argument in that thread which is basically pro-vote and pro-reform has 51 votes, the next top parent comment takes a fairly neutral "i like this experimentation" line and puts forward an alternative case of voting/labeling and has 35 votes, and the first properly against-seeing-votes parent comment has 23 votes. obviously not all of those are endorsements of the idea expressed, of course, but that's a decent drop off between ideas to say the least, one which was entirely erased by not being able to see votes.
I think that would be the case regardless of vote count visibility.
Also, thank god. I must know how (un)popular I am.
yeah. i don't think this is really something that can ever be completely eliminated, because by nature any comment which has been up longer than another post will have more time to collect votes than a post that's been up shorter and if all things between them are equal, the older one will almost inevitably end up having more votes than the younger one purely because of the time difference. that said, i don't think it's exactly an issue on here as much as people sometimes say it is; i've seen plenty of older posts become some of the top-voted posts in a thread and you need look only to the announcement of the change for one such example, since most of the top ten by relevance were not among the first ten posts in the thread.
It can be greatly diminished by using a random sorting for comments, so older comments get presented randomly throughout the thread alongside newer comments, rather than sitting at the top because they got a few early votes.
i really don't care much for the idea of random sort, to be honest with you. it can be implemented well but i find it irritating to browse in general outside of certain contexts where it makes sense like QnA formats and contest modes, and i would probably immediately switch back to sorting by relevance if it became the default on here because i just don't see this whole vote thing as a particularly big issue.
I wasn't recommending a random sort. I was merely identifying it as a way of reducing the effects of "first in, best voted".
That said... I wouldn't object to a random sorting (while retaining visible vote counts). It would reduce bandwagoning, and would give equal visibility to all points of views - even the controversial ones.
While this is not as grave as on Reddit due to the options to highlight new comments and to collapse already viewed ones, I would suggest that users could change their default comment sorting method (e.g., I would default to Newest First).
To spin out another idea, it might be an interesting experiment on tildes to use the sorting we have now, except put one new comment at the top. Perhaps the newest one, perhaps a random one of the last n comments.
Not yet, though IIRC there is a GitLab issue for that
Ninja edit: Issue 153
I wonder if tweaking vote weight based on age would help. Give new comments some sort of boost in effective vote count for sorting based on the age of the topic at time of comment, the age of the comment, and max real vote count. A fresh comment would get a big boost, but the boost would diminish as time wears on and it'll increasingly rely on its real vote count. Doesn't have to be a linear drop off either. The result is a hybrid between newest and most popular sort, allowing the most popular to generally rise to the top while giving fresh comments the opportunity to be seen and evaluated.
The default sorting method on Reddit,
best, takes time into account to prevent that. It's not uncommon for newer comments with fewer votes to come before older comments with more votes.So I think the primary change I felt from not seeing vote totals was a general sense of doubt as to whether anyone is actually reading what I'm writing or if I'm just talking to myself. People are way more likely to vote than to comment, so any post where nobody replied it felt kind of like "Did I spend all that time writing something only to have nobody read it?" It definitely felt a bit like I was shouting into a void at points.
Now that I see them, it seems like all this time I've been getting more votes on comments than ever. I think maybe a 24 hour lock might be fine. I didn't see any impact on how I processed other people's posts.
I felt more of an obligation to reply to people, which I didn't like. Whereas with votes I might just like a reply if I didn't have much to add, when votes weren't visible I felt bad not having a way to be like "hey I read this." (Of course, no one knows I voted for their reply, so maybe it feels the same to them...)
I don't know if it was just my personal withdrawl from precious virtual internet points that was clouding my perceptions, but I got the feeling that the conversation got disjointed when votes got taken away. It seemed that more people wanted to express their minute changes and there were a bunch of almost-but-not-quite redundant posts and comments. Though it also did seem to get better towards the end.
you're not alone in that perception. for example, i feel like the gigantic announcement thread could have easily been fifty or a hundred comments shorter (and orders of magnitude more manageable) with votes than it actually turned out to be purely because of repetition that voting might have eliminated.
I'm not against vote counts, but I was totally fine without them. It would be nice to see it on our own comments (a la HN).
With reddit and other communities, I tend to stick to smaller groups that are more 'I have a question, does anyone have an answer?' -- so I'm not accustomed to getting a lot of karma / points / response off of posts anyway.
I also view points purely as a way to vote up more meaningful, quality comments -- and nothing more, so on a personal level the points are meaningless.
It was a fun experiment. It was interesting to see the different approaches to karma / voting. I've always assumed it was purely two camps (agreement vs quality), but the last main thread really broke that down.
Annoyingly, this week coincided with a period where I spent less time on Tildes, so I wasn't as elbow-deep in this experiment as I would have liked to have been.
That said, the few times I did comment, I ended up with a sense of shouting into the void. When noone replies, and (effectively) noone votes, it feels like I'm just writing for my own benefit. If I wanted that, I would type my thoughts into a text document on my computer. I post on Tildes and other social media for interaction. Even if people don't actually reply to my comments, they can vote on them. This provides confirmation that I'm being heard.
Some people have talked about how they get guided (for better or worse) by vote counts on other people's comments. I don't understand that. The counts are shown at the bottom of each comment, and they're very low-key. They're practically invisible even though they are displayed. And, even when I do notice them, I don't really let them guide me. I know what I like. I don't need to see a vote count of (1) or (20) to tell me whether a comment is good or not; that's up to me to decide. I've seen high-voted comments that I think are merely ordinary, and low-voted comments that I think are gems - and I'll vote accordingly.
That said, it can be helpful to see just how popular or well-regarded a comment is. Is it sitting at the top of the thread because it has only 8 votes in a low-voting thread, or does it have 50+ votes and has outstripped the other comments by dozens of votes? That sense of knowing how the community is responding to opinions is also a valid part of participating in social media.
I honestly forgot by Sunday that vote counts were hidden. It wasn't until the other ~tildes.official post that I remembered it happened as was a temporary thing. Looking back, I think I maybe I got one or two more comments on my posts than normal? I didn't notice any difference at all really. I think no votes makes more sense and is a better design to me personally but apparently people really rely on votes to read threads so ¯\_(ツ)_/¯ guess I'm the outlier.
I kind of miss not knowing. But also see the value, tough decisions.
Maybe have the vote counts hidden for a first time viewing, when seeing the comment for a second time; this triggers the vote to be displayed. Using the same system as highlighted comments.
It kind of treats vote counts like a movie spoiler, but once you've seen the comment it doesn't matter.
I thought of that, but concluded you can't really hid this information anyway, if people want it they will always find workarounds, most likely in extensions. My point being, it's best to design the feature so it's the most useful for those who actually want it, rather then compromising a solution just to make it harder to circumvent.
That said I kind of like the idea of having votes only visible after a vote has been cast, makes ~ distinct from other vote based link sharing sites out there.
On the announcement I declared that I'd stop voting for the duration of the experiment. At this point I don't feel an incentive or need to resume. Not being able to give single-click anonymous kudos or to support good comments was a problem at first, but I got used to it pretty quickly.
I can say that I feel like my comments got slightly less votes than they generally get, but it is impossible to objectively prove that. I also don't think I interacted less frequently than I normally do, and I did not feel like overall activity dropped.
Ultimately, I'm in team no-votes, not having them does not really bother me, and not evaluating each comment as worthy of a vote or not was nice. I'll also change my mind on the private replied / whisper comments idea: for the duration of the experiment, I did not really feel a need for it.
I also want to reiterate this thought: if we will have votes, I prefer them to be visible; if we won't have votes, we should remove them entirely. The reason being in between the two we're trading off transparency.
What would Tildes be without votes? A collection of threads in chronological order, or maybe even random? I don't think it would be very interesting.
That idea is surely valid (tho I find myself in the neutral, maybe better without votes camp ATM), but what I meant was that if we won't display votes, we'd rather remove them entirely, because that hurts the transparency of the platform.
I agree.
I'm personally glad to see the vote counts come back, tildes felt a little barren without them.
I wasn't here last week, but I think I'd quite like not being able to see vote counts. They influence my view of a comment before I've had a chance to read it, and I'd like to give all views equal thought.
Although, having the vote count at the bottom of a comment as it is now is a good compromise if hiding them completely isn't the way forward.
One of the biggest issues I see repeated is that without votes, people don't know if their comment is being seen. What if we started tracking comment view counts based on page load (maybe even broken down to pre and post edit(s))? Privacy policy makes things a bit difficult for accuracy though. Could use general page loads, but those are not unique views, so less accurate and easily gamed, but could still be useful if shown only to the OP. I don't know if it's possible toto hash an internal user id and a comment id together in some way such that unique views can be tracked without easily being traced back to the user (if having to brute force user+comment id combos is not private enough)
If voting is invisible, it'll give OP some idea that they're at least being seen. If voting is visible, it could be interesting information for OP to evaluate; 1 vote on a comment only seen a handful of times is a lot less useful feedback than 1 vote on a comment seen a hundred times. A similar setup could be extended to topics for similar feedback.
Beyond OP, such info could be useful in some capacity internally. For example, flagging potential spammers if a user posts a lot of topics that get a lot of views but low votes (enticing or misleading titles but low quality actual content), or at least weighting topic count less for activity in the trust system
That doesn't solve the issue on large comment threads though. View count for the topic doesn't indicate your comment has been seen in any way, especially when it might be buried amongst a bunch of others. And there is probably no way you could possible hope to track that without monitoring scroll position on the comment page.
Similarly, particularly in a large thread, is a vote count low because people saw it and don't agree (or don't think it's a good comment) or because people aren't seeing it? Absent additional information or very high vote counts, votes are a poor indicator of whether your comment is being widely read, but feedback on simply being seen has been a major objection to removing visible votes. Perhaps the underlying objection is the loss of active (pressing vote) support, but is being described as wanting to know that they are being seen?
I don't think it's about people trying to determine if their comment is being "widely read", they just want to know that it's being read at all and appreciated in any way, but without visible vote counts you can't tell that is the case unless someone replies to you. It can feel like talking into the void without any feedback, and even something as trivial as a vote can help alleviate that.
And I think it's pretty obvious to nearly everyone that has used sites with similar comment systems that when you make a comment late into a topic's life, especially in one already populated with lots of other comments, that a low vote count is largely due to that and not because nobody agrees with you or appreciates your comment. It's all relative, and people learn to temper their expectations based on those factors.
E.g. If I make a reply in a deeply nested thread in a topic with 100 comments already and get 2 votes, that satisfies me just as much as if I made the very first comment and got 50.
But the problem with view count is that while in the very beginning it may give you a pretty good indication of how many people actually read your comment, it gets increasingly inaccurate the longer topics go and the larger the comment section gets. And as the site scales up that window of accuracy is only going to shorten, since the comment volume will increase over time. So it ultimately isn't that useful as an indicator of whether your comment has been read or not, and is completely useless for determining if your comment was appreciated at all, unlike votes which can serve both purposes adequately.
To me, knowing a comment is being read and being appreciated are two separate issues and can be handled separately. I've seen many people brought up worrying whether their posts are being read or even seen at all, before even touching on community reaction to the contents. In general, the worry that a comment on a social media topic might not be read at least a few times in its lifetime without explicit feedback is absurd to me. To that end, I see one or two votes on a comment no different than assuming at least a minimum number views based on probability as page views increases. Wanting feedback on how a comment is received by the community is a different matter, and I think can be tackled better by something other than plain votes (there's been many other approaches discussed, such as labels)
I don't disagree with you there are probably better ways to do it that should be explored... but I definitely don't think showing view count is the way to go about it or an improvement, is all I am saying. But for now, voting and showing vote count serves both purposes (and more) adequately, and hiding the count is throwing the baby out with the bathwater.
I felt disengaged. Kinda like I feel about Hacker News: more a website that I seek for content than a social network. Commenting is too much effort for lots of people and voting provides a sense of community that was greatly diminished in the last few days.
I got more votes during this experiment than I have before here. Not sure what it means, but it was a nice surprise