18
votes
How do you tag comments?
I see in the documentation that we are supposed to be able to tag comments, but I see no way to do that. Is this feature yet to be implemented?
I see in the documentation that we are supposed to be able to tag comments, but I see no way to do that. Is this feature yet to be implemented?
The code is there. It's off for now since some people think tagging 'troll' is a substitute for downvoting, even when the person they are tagging is clearly not a troll. We need to take a step back and go for a different approach of what tags to put up and what they do before turning it back on again.
I think an important part of Tildes is only having the ability to vote for a topic or comment and not against it. Personally I think when tags come back, they should reflect that and only be tags that are positive.
I've been thinking along those lines myself... the negative ones are likely to be abused, though we might be able to diffuse that with humor - someone made a nice post about how they did that in their discord channels and it seemed to work (if I can ever find it here again). If we stick to positive and utility tags we might get better results.
One example I was thinking of is an 'offtopic' tag. Enough people tag a comment as off-topic, it'll show up as collapsed in the default view of the comments. It's not punishing or removing anything, just collapsing parts of the threads, exactly like we'd discussed a couple days ago for already-read comments on a second visit. There's no way to know how that'll play out, though. I can certainly see people annoyed with certain viewpoints trying to auto-collapse them in retaliation.
I think maybe there needs to be a delay between when someone joins tildes and when they get tagging capability... so they can observe them in the wild here and let the cultural norms sink in. Low hanging fruit of the trust system perhaps. Most people just coming here from other sites are still going to be reflexively thinking the same way they do on those other sites. It may simply take some time to acclimate.
I don't understand why all this overthinking about tags when it was said already several time that tags were hidden because they should work on a "critical mass" system.
Am I remembering it wrong?
The way i remember from the last discussion was that a tag should begin to be visible when a certain threshold of users have used it.
Now then, my personal suggestion was to make the threshold a percentile of the partecipant on the topic, because maybe a topic (or group) has less users partecipating and if the threshold was a fixed number, they would never kicks in. But that could be a following iteration of the system.
Also, the trust system should have a weight in that so that the trust of a user should act as a multiplier for how a user tag weight in reaching the threshold, giving that it should be capped to 0.49 tops, so that it take at least 2 highly trusted user and some to actually have a tag appear.
I think that might be a part of the solution. If the tags aren't visible until they activate it should stop the bandwagon effect.
Once they do activate, we might be able to implement some sort of disagreement metric, so people can give feedback on how the tag was used, maybe even appeal it.
The colors we're using now for tags are a bit in-your-face, bright and distracting. Maybe we need something more subdued and boring.
That will be impossible to moderate. Given any set of possible tags, people will always be able to use one in a negative manner. It's basically another incarnation of the Scunthorpe Problem.
Wow, so this phenomenon has a name. I used to laugh whenever we got false positives in my music subreddit when Automod found the word "semen" in "basement". I can't remember how, but someone adjusted the code to flag only standalone words.
Pretty sure that was me on l2t :P
It's fairly simple with regex, either search for whole words r'\b(semen)\b', or (and more accurately) do a negative look-behind for 'base', r'(?<!ba)semen'
It was a basic clone of slashdot, but with one key difference - on slashdot, the people assigning the tags/moderating were not part of the discussion. Instead they were coming in through the meta-moderation interface, and so were less likely to have any reason to bias their judgement.
Perhaps we can take a cue from that. What would happen if you can't tag comments in threads where you make a reply? Leave tagging to the lurkers, basically. There are a hell of a lot more lurkers out there than commenters. Most sites leave lurkers completely out of the equation. Perhaps we should show them some love and see what happens.
For starters, you make comment-tagging available only to trusted users within a group. Suddenly, you've killed off most of the drive-by tagging-as-disagreement.
Also, in the hierarchy of trusted users / moderators, there will be higher-up moderators who can see how lower-down moderators are using and abusing comment tags. If someone continually abuses them, they can be demoted down to a user/mod-level that doesn't have access to comment tags.
It's not impossible to moderate this at all.
IMHO, better group moderation tools are almost a pre-requisite for the tags to be re-enabled so that tag abuse can be handled.
Also, I would suggest that the tag style should depend on the number of those who used the tag (e.g. if only one user tagged a comment as off-topic, then it is shown in plain colour, if there are ten, then it is displayed as in the example below, and if there are one hundred, then it can be displayed in the "heading" style—this all unless overridden by the group's CSS).
An example of a thread with tagged comments from the ancient times
Hahahaha. I remember thinking "these tags are ripe for abuse" when that happened.
I loved that whole thread. Makes me smile just thinking back on it.
What were your thoughts, Eight? Did you get a kick out of people poking fun, or was it annoying that you had just said you didn't like something and people started doing it?
It really didn't bother me. I kind of set myself up for a bit of mockery, and if we can't laugh at ourselves then we can't justify laughing at anything or anyone else.
I'm glad!
I've seen that thread before, and it really makes me think I wouldn't want to be part of a discussion forum that uses those tags. They are more ripe for abuse than for fostering quality discussion.
Jokemight be useful for weeding those out, butNoise? That can be used to downplay a person or argument you disagree with regardless of what they posted, andTrollis just a straight-up insult. Are we here for that? Or for actually talking to each other in a respectful, mature way? I hope it's the latter, and that we don't drive potential quality members from posting here because they see people with good points and sourced arguments buried underTroll x6orNoise x10. Let the votes decide what's quality and what isn't.I share these concerns, that is why I wrote the part above the horizontal line in my comment above (penalties for tag abuse, less visibility for tags unless added by sufficiently many users).
I'd like to read the arguments for tagging comments, which are probably here somewhere, but it's currently difficult to search the site. Is it so you could potentially weed out things tagged
JokeorTroll, for example? Because then we'd need a qualifier, such that oneJoketag doesn't necessarily filter that comment out, but more would, with a set limit. Then it gets muddy, with a comment with oneNoisebut fiveJokeand oneTrolltags being filtered because of the fiveJoketags...it becomes a bit unwieldy.Group mod tools would help, but I still think it makes us look like a bunch of judgemental people, and prejudices opinions of writings without great benefit other than filtering out particular tags, which as stated above gets a bit messy. How likely are we to give a proper fair read to a long post with
Noise x5tagged at the top, without first forming the opinion that it's most likely noise? Not very.Clearly that was a suboptimal set of tags to start out with. Hey, at least it was rolled back immediately and we're talking about how to make it work, which makes this place infinitely better than most forums in my mind.
The argument for the tags is letting them be a crowdsourced form of moderation. We all know that threads need moderating, more and more of it the larger and noisier they get, or the more heated the debate becomes. That moderation usually falls on a handful of people and they burn out and/or use poor automation to cope. That's not a solution.
Outsourcing most of the moderation to the users, however, might be a solution. The question is, how the hell do we do that exactly?
The trust system should gate access to those tools and lessen abuse by limiting the use of that moderation to people who are participating in good faith - people who have been around a group for at least a little while and have absorbed that group's cultural norms. In theory, anyway. That could still be a colossal bust too, we won't know until we try.
Presuming we can gate access successfully, we can say that we have identified the users we wish to vest with expanded powers. The question now becomes, what powers shall we grant them?
The comment tags are one form of 'power' that can be widely crowdsourced here. The tags provide cues to the system in a very basic way. Tag X has function Y once Z people have tagged the comment and met the threshold for that tag (which may vary group to group, set by mods, because big groups will need different thresholds than small groups).
That's all it is. A community vote on a function to be applied to the comment. It's presented as a tag to keep it simple in the interface. It has a threshold for triggering to make sure there's at least some consensus among those doing the tagging.
So what tags do we want, and what will they do? How do we keep them from being abused in the heat of the moment?
What could tags do? Just some examples...
This might be an overly simplistic view of tag moderation. Even slashdot's system is old/primitive, there hasn't been a lot of advancement in this area, most sites don't have comment tagging. If we're going to truly crowdsource moderation to the userbase here, I think we're going to need to evolve this concept into something more clever/nuanced to do it. I don't know what that is. Hopefully someone reading this has ideas and will share them. ;)
I think it's important to make the distinction between comment tags and submission tags. Those are very different use cases despite the similar nature of the system. Submission tags are for sorting and identifying content, it's a classification system at its core. Comment tags are to help the discussion self-regulate, so that the mods/curators don't have to spend every waking moment dealing with the comment moderation - or ignoring it, as is the case on most of reddit, due to the workload being unmanageable.
Does that help explain why we're talking about using a comment tagging system here?
It does help explain it, yes. Thank you for taking the time. If you don't mind, a few points:
Thus they could be invisible to the no-access user, correct? I really think the only way tags work effectively is if they are invisible to almost every reader who might wish to join the discussion, otherwise prejudice is going to occur, without question. There's no way to approach a comment with a visible
Noisetag without pre-judging it; it's simple human nature to assume it must be noise, since it's tagged that way. A reader doesn't know the predilections of the person who tagged it. It might be "noise" to them; to another reader, it might answer a question.If instead, as you say, someone vested with powers could tag it and it would have some effect, such as being hidden after a certain threshold—although that can get messy for child comments of a quality top-level comment—that could be avoided. (As a note, I hate this idea on grounds of censorship of non-offensive content).
"It's presented as a tag" is my problem here. Are you seeing the prejudice in a system of visible tags? Even if certain vetted users are permitted to tag comments, they have their own biases, and they'd be causing readers to pre-judge comments before reading them. If the system used invisible tags, I could still use them to filter my comment preferences (I probably wouldn't, but some might) without knowing "hey, a guy who's been here a long time thinks this comment is
NoiseorOff-topic, I might as well skip it."Sorry to keep going on about this, but reading that older thread with the tags lowered my expectations for this site a little. I'm enjoying using this site as it currently is, and I have very little interest in participating in a site that looks anything like that older thread.
The tags could be invisible, or at least invisible to most users, yes. Their whole purpose is to give cues to the system, not visual cues to the users - though we need them to appear in some form so people can click on them to use them. Perhaps under the comment like vote and reply.
We need to do something, or else comment threads here will turn into reddit (well, not quite that bad without downvoting). That's the 'natural' state of a busy thread unless something is actively done about it.
Agreed.
This gives me hope.
Yikes, I really would not want a tildes thread to look like that with tags. It seems like it would take away from our most shared goal which is high quality content and discussion.
It used to be, but was rescinded after a period of abuse. I think it's planned to return though, after some reworking.
That's good to know. I couldn't seem to find a named issue on the issue tracker for it.
It used to be enabled, but has since been disabled.