Slashdot Mirror
Ask Slashdot: Going Beyond Comment Threads?
asa writes "The Knight Foundation and Mozilla are running a series of news innovation challenges. The goal: get the world's smartest hackers thinking about how news organizations can harness the open web. The current challenge is all about comment threads. This seems like the perfect question to pose to Slashdotters: how would you foster more dynamic spaces for online news discussion? How would you preserve the context of online discussions and stamp out trolls? All ideas, technical, practical or impractical are welcome. What technologies (federation, atomic commenting, moderation, algorithms) would you employ? What are the immutable social dynamics? Knight and Mozilla will work with the best challenge entrants to deploy the solutions in newsrooms at Al Jazeera English, the BBC, boston.com, The Guardian, and Zeit Online. Submissions are open until May 22nd."
People can say whatever they want, moderation points stamp out trolls and assign relative values to posts (not always the best system, but not bad), etc. Of course, for it to work, you have to assemble a pretty smart/knowledgeable/literate bunch of people (that's the real trick). And you would still have to avoid kdawson stories, of course. Not a perfect system by a longshot, but one of the best.
Most of the systems I've seen on news sites for commenting have ranged from "suck ass" to "MAJORLY suck ass." Moderators are either too tough (nothing controversial gets through) of too lenient (leading to comment threads loaded with spam). Just go look at the "Wired" story comment sections sometime. Half of them don't work at all, the other half are loaded with spam, and some of their stories don't seem to let you comment at all. And that's from a *tech* magazine.
SJW: Someone who has run out of real oppression, and has to fake it.
Full identification like Facebook, moderated semi-anonymity like slashdot, full-on unfiltered anonymous chatter. Each has fans and faults.
So make three tabs and call it a day.
I mean maybe I'm missing something, but is there a rule that there has to be one best way?
Help stamp out iliturcy.
If I came up with the killer idea in this area do you think i'd tell you?
Everyone has a right to speak, even idiots. If you don't like what they have to say, then just add them to your ignore list. Trolls/idiots polluting forums is preferable to censorship like happens on some boards (Sony erasing negative posts about the hacking).
Other ideas:
- no point system or post tally. People don't deserve to get points just because they post a lot. People don't deserve to get points at all, for the mere act of expressing an opinion.
- threading is essential, so the replies are tied to the original post
- keep it simple. Plain text. Uses less bandwidth.
My AC stalker: " I personally agree with your posts most of the time, but that won't keep me from modding you troll"
How would you preserve the context of online discussions and stamp out trolls?
Punching them in the face over TCP/IP.
Well, you did say impractical ideas were welcome!
Make it impossible to post a comment without having first RFA.
Many usenet readers provided an astonishing number of features that we have lost in the move to web forums. Maybe something was gained, but much was also lost.
Killfiles were useful to stamp out trolls. These days there could be a feature similar to adblock subscriptions that would block known trolls across all forums.
The reader features themselves on the better clients were MILES more advanced than what is done today in web browsers, even with AJAX. A web browser is a good tool, but it isn't the right tool for everything. It doesn't seem like the right tool for large scale discussion forums, although it can "suffice" for them - it just isn't as good as a dedicated application. Also with a dedicated app, you get your choice of which one to use. With a web forum, you get whatever the forum software gives you to. It takes choice away from the user and places it with the site.
Trolls only. I've been trying to post on presstv.ir for a couple of years, but their moderators won't allow any sane or reasonable views to taint their pseudo-islamic purity.
They feared that it could be used to suppress protest or support unpopular rule.
If you're asking slashdot, then the obvious answer world be "with slashcode"
I've always thought that assigning negatives and postivies to Slashdot mod categories was limited. I'd like to see mutiple axes of moderation, with the axes you want to view left up to the users.
For example, I might really enjoy viewing (-2*Troll+Funny) as opposed to the default (Funny-Troll).
For all intensive purposes, "whom" is no longer a word. That begs the question, "who cares"?
Ask Thunderbird developers: Usenet has all of it: threaded discussions, personal bozo bins and if you prefer a more detailed approach to outright plonking you can score articles. Add to that a Bayesian spam filter and what else do you need? The only thing on the web that comes close to it is /.
Oh well, Usenet is decentralized too and it now has great retention, great local search, etc, etc... I guess sometimes innovation is copying something that's already been out there for like 30 years.
Where slashdot fails:
a) Anonymous Cowards are seldom read, and seldom moderated up.
b) the Javascript filtering makes it impossible to search.
c) Mandatory login to get rid of the 50 comment limit.
Where Twitter and Discq.us have fared better:
1. Universal login, so I don't have to keep creating accounts on 10,000 damned sites, blogs, and everything with a comment field.
2. Disqus - "Likes" somewhat like a karma system.
3. Twitter - Followers, eg follow those that contribute, so by this nature it "whitelists" when they comment, and hides all other comments by default.
Where they've failed:
1. Any jackass can create a twitter account, this is the number one problem, verifying identity, while keeping off trolls and sockpuppet/meatpuppet/astroturfers. We don't need to verify WHO they are, just that they are only ONE UNIQUE person. The easiest way to do this is by having state/national ID database that is queried upon creation of facebook, twitter where a [x] This is my true name, box is ticked. If the box is not ticked, they are defaulted to greylisted (eg slashdot anonymous coward) and have to rely on the karma system alone. If the box is ticked, then the karma system no longer applies and instead a civility meter is kept by twitter/facebook, too many people complaining about that person's behavior will push them into greylisted and down to blacklisted if they're being obnoxious everywhere. Once they're blacklisted, the site has to explicitly allow the person immunity from the blacklist (eg put them on a whitelist) to allow them, otherwise they're effectively silenced.
2. Search Spam. Twitter's search function tends to show more spam than anything useful. This could be better effectively if they simply "end-tailed" all the links being posted and eliminate links or entire spammer "teamfollowback"'s by looking for the same link being posted by people with no posting karma. 250,000 followers and only one comment, yes that's a spammer.
Open discussions can either be primarily informative with an "anything goes from anyone" approach, or primarily educational in that some attempt is made to improve the signal-to-noise ratio by dealing with the trolls and spammers. In either case, any given forum will have to initially take a perspective to be either informational with the exercise left to the reader for picking out the wheat from the chaff; or if their perspective is educational and hence doing some guiding of the discussion for relevance to the topic. And clearly, some venues will have a more obvious beneficial choice to make than others. This is the internet, and I would hope by now that we can get over the false notion that there is a one-size-fits-all solution for anything.
There are still discussions on technical topics, and trolls are relatively tame in those newsgroups. Usenet nodes almost universally use spam filters now, which helps a lot. Usenet is not dead and there is no reason to refer to it in the past tense.
Palm trees and 8
Add a "Crickets" button similar to a "Like" or "+1" button; the count is indicated next to the post.
Pushing it is the equivalent of saying "Not going to dignify this with a response." The idea is that the troll has gotten up on stage, spewed a bunch of crap into a microphone, and 5000+ people have utterly ignored it ... the only sound is that of crickets chirping.
Usually whenever the news has obvious political implications, many posts will be repetitive advocacy for one side or the other, often laden with mudslinging and sarcasm directed towards the other side. This may or may not be useful as an emotional outlet, but it's hardly enlightening. Unfortunately it's hard to get around that without heavy-handed moderation.
I think a facebook twitter combo would be a pretty good solution. On a given newsfeed you should be able to see the comments of those who you are following. You would obtain those friends through a one sided relationship like twitter (i.e. they don't need to grant you permission) and you could also see the posts of those who belong to specific groups (i.e. I hate Republicans or I hate Democrats) -> that way, you could find more individuals to friend. You could perhaps also have an option to see posts from random individuals who have a high degree of popularity (i.e. many followers) so as to spice things up a bit. The ignore feature would be just as important, you could eliminate certain folks from your newsfeed or certain groups. The ignore would override any setting for group following.
It's pretty common sense and allows people to see what they want and from whom they want so I'm not sure why this has not been done. It's computationally and architecturally a bit demanding, but not much more so than facebook itself.
The downside is that you'll enable people to wall themselves off from the viewpoints of others, so there will be a lot of in-group discussions without outsiders telling them they are wrong. But this is human nature, already happens for the most part and really isn't of concern to a site that just wants to generate traffic.
Can anyone poke holes in my idea? If so, I'd actually like to hear it. I've wanted to see this type of thing for a while.
If you can read this... 01110101 01110010 00100000 01100001 00100000 01100111 01100101 01100101 01101011
Let's ask /. this question:
How come after year, I mean it's been more than a decade for me here for sure (if I remember right), how come after all these years this site hasn't figured out the simple things about comment threads?
COMMENT THREADS!
Look at this fiasco - that's a comment about somebody leaving yet another comment in that thread, and quite a number of comments there are the same, redundant stuff, and why?
Because /. can't do comments threads right.
Of-course you can disable this crap on /.
(Account, Discussions, Classic Discussion System (D1))
and
(Options, Simple Design)
but a story on /. about a better design for comments threads - now that's irony.
So to the question:
How would you preserve the context of online discussions and stamp out trolls?
- I will say: forget trolls. Get the basics working and don't screw it up first of all - let the people SEE what the thread is first of all. Don't hide comments in threads by default.
As to trolls, etc: have simple "like/don't like" and have thresholds, nothing else should really be done. You can't get rid of trolls, and look what /. is doing and don't do it.
Do NOT force people to log out and post as ACs if they rich some weird 'threshold' (number of comments they can leave under their user name per 24 hours) - what good does that do? People register other accounts or they post as ACs. This is NOT good for discussions.
Do you have a discussion forum or is it a chat room (IRC like)? I think that's the first and only really important question. Do you want to keep history of all the comments or not?
Here is what /. is really doing that's pretty stupid: not showing the entire history of comments for non-AC users. As stories age, they disappear, contexts disappear. What's the point of having any history on line if it's uselessly unsearchable? There is no index.
There is no way really to link to an older discussion that maybe of some value.
Also for various political reasons on this site, comments are often moderated high up, and then after a while they are moderated down only so that people wouldn't be noticing them, even if they are totally pertinent to discussion, no trolling, no flame, those are just unpopular views and a coordinated moderation attack pushes them down where nobody is reading.
Don't allow political dissent to be drowned on your site by shills and just by those who don't like what you have to say. Have the "like/don't lie" feature - that's useful. All this other nonsense is just counterproductive if you don't want to run a site, that's dominated by one single mindset.
You can't handle the truth.
I've always thought that allowing users to moderate comments is a good idea, but it seems that a lot of people either don't know how, or think it's a waste of time. If users were exposed to the merits of moderating comments and the good that can come from it, it could be a good venue for filtering flaming and trolling. However, by that same token, users could simply utilize other users to downvote/rate other comments because of a disagreement, argument, or whatever. It'd be nice to find a way to prevent this while still allowing users to do comment moderation.
"...the number of UNIX installations has grown to 10, with more expected..." - Dennis Ritchie/Ken Thompson, 1972
So what does "harness the open web" mean? To the news organizations it probably means "make money". Comments are largely irrelevant for that. If news organizations want to attract readers (assuming that translates to "make money") they need to do two things - 1) report news that people feel is interesting or relevant. and 2) provide intelligent analysis of the news (not discussion forums). They could use a forum to generate ideas for further investigation/reporting/answering tough questions that the media doesn't do too often. Of course by the time they ask questions and do some further investigation the story will be a bit old. And this also doesn't work with stories about tornadoes and celebrities.
Or how about we don't! The new organizations can do there own dirty work. If someone could harness the power of the web, in a manor that helps the media, this would have value, it would potentially make someone very rich. Media has its own agenda, it isn't in our favour (as in the public) the worlds best hackers should be doing what they have always done.
News is about trolling for eyeballs. The most sensational, shocking, scandalous and salacious stories attract the most eyeballs which means more advertising revenue. If it bleeds it leads. How can trolls be stamped out when the news media culture is rooted in a form of trolling?
atmosphere infactdead, replaced with atmostfear.us (Score:mynutswon; left to our own devices)
once one lie is 'infactated', the rest becomes just more errant fatal history.
disarm. tell the truth. the sky is not ours to toy with after all?
you call this 'weather'? what with real history racing up to correct itself, while the holycostal life0ciders continually attempt to rewrite it, fortunately, there's still only one version of the truth, & it's usually not a long story, or a confusing multiple choice fear raising event.
disarmament is taking place based on the pure intentions of the majority of the planet's chosen to be depopulated, population. as the biblical fiction based chosen ones have only one ability, which is destruction for personal gain, they just don't fit in with all the new life extending stuff that's we're being advised to ignore. life likes to continue, advance etc... deception & death appear to have similar ambitions. with malestromous monday on the horizon, wouldn't this be a great time to investigate the genuine native elders social & political leadership initiative, which includes genuine history as put forth in the teepeeleaks etchings. the natives still have no words in their language to describe the events following their 'discovery' by us, way back when. they do advise that it's happening again.
see you at the million babys+ play-dates, conscience arisings, photon gatherings, georgia stone editing(s) etc...?
Well said. Best example is how any pro-apple comments are moderated to infinity and anything against apple is modded down to hell - irrespective of facts and data.
Slashdot system works well in general though - just not for apple threads.
So many sites have very poor verification that a post has been submitted. If user 'x' submits post 'y' and then submits post 'z' with the same verbiage, would be nice for the system to see the duplicate post and not post it. Would also like to have the ability to delete a post/thread if I was the one that posted it. Those two changes would cut down dramatically on the number of bogus posts in my opinion.
W/out trolling there is almost no reason to even read the comments..
Trolls may be a problem. Personnally, I think that special interests groups present a more formidable challenge.
Once they target a thread on an organized attack, either by themselves or through a PR agency too happy to cater to their needs, there is little that a few, by definition disorganized, moderators can do. The tone heats up in minutes, you can see that any of the seasoned intelligent commentators stay away from such threads. Sometimes, they back off from the site entirely.
Slashdot has a pretty impressive record, and the administrators surely have valuable experience in this regard. Even then, from time to time, you see the sturdy moderating system collapse under an persistent assault. This is always a disheartening experience for me, to see bullies have their... I mean our, cake.
In these times, I always wonder what we could do to prevent this from taking place. I do think that an awful lot is at stake: public interest, to say it in two words.
After thinking about this for all of one minute, I thought that an interesting way to go would be to separate a single large community of readers into smaller communities. It would be like going from a message board posted in the kitchen of a massive office, to lots of little, different message boards at each water cooler. You would often see the same people, thus forming a less anonymous message board. Just feels this would mimic real life a little more closely.
Why not take it a level further and allow commenting not just on news stories, but on any desired url?
A comment-system could even be integrated in the browser. Imagine just opening a webpage, clicking the "comments" button, and seeing a bunch of moderated comments (perhaps even in slashdot style). Now that would be awesome!
If Pandora's box is destined to be opened, *I* want to be the one to open it.
Actually an interesting point. Maybe some webservice could be implemented as a substitute to the vanilla browser capability? That'd be a great idea and one that I've at least never considered.
If you can read this... 01110101 01110010 00100000 01100001 00100000 01100111 01100101 01100101 01101011
what I'd like is a comment system where the comments are shared across sites.
If a popular news item comes up, I might be interested in the editorial slant at theregister, but perhaps I first saw it on digg, and there's a tech angle at slashdot that's interesting too. It's the same *story* and I want different news outlets to do their own research, conclusions, and editorializing, but it would be cool if the comments were shared so that if someone posted a witty +5 remark on slashdot, people on digg would see it and could respond there, even if they don't have a slashdot account.
You have to build positive karma to earn downmod points. Still limit how much they can be employed but require positive contributions to earn the ability to downmod. Too often we see sockpuppet accounts downmod unpopular opinions or mark them as troll simply because they disagree. There's no disincentive if these sockpuppet accounts accrue modpoints simply by sitting around.
Kwisatz Haderach
Sell the spice to CHOAM
This Mahdi took Shaddam's Throne
We enjoy the slashdot system because we don't mind rehashing the same discussions every time a new story comes up, but it's more than a little redundant.
I would detach comment threads from the stories and tag them; old comment threads could get automatically attached to a new story that's similarly tagged, and then moderated down if it's no longer applicable or up if there's nothing new to say; presumably, some sort of pruning or metamoderation to cut it down to only the best posts. In theory, wise posts could actually have f--king staying power instead of being one-offs on the latest sensationalist headline that have to get rewritten every time a site moderator reposts f--king bullshit that's not actually news until everyone gets tired of writing it, and we could stop pretending that repeating the same damn truths as last week makes you 'insightful,' 'interesting', or 'informative.'
How about everyone getting to choose their own way of tagging and displaying every comment and user with an optional added numeric modifier for every tag?
Some days I might want to see (or hide) for example what the most popular "+3 Constipated"(and up) comments from anyone modded at least "+2 United-Fruit-apologist" by self described "Anarcho-Marshmellowians". At other times I might choose something less ridiculous, involving tags like "Conservative", "Insightful" and the like.
One could also choose to view comments in the style of reddit or slashdot (except maybe everyone would always have points, so the slashdot style would be filtered by mostly most popular moderators calculated in some way.)
Some of my favourite people are from th US; Vonnegut, Chomsky, Bill Hicks.
If the forum is open to everyone and it's about politics, then you are going to deal with a lot of trolls and zealots and astroturf operations. You cannot avoid that, but for my money, just letting it be what it is and having some type of "report spam" button is the way to go. Cut it off after some number of days or hours hrs. Oh, and allow anonymous comments and make sure it's threaded.
look sig is kool
Slashdot is one of the best I've seen. If you allow anyone to be a moderator, and have unlimited points past 5, you get all sorts of people gaming the system... And the hivemind. Besides those systems were no doubt influenced by Slashdot
I've thought of a superior system to Digg.com, where instead of everyone getting upvote/downvote priviledges on a global pool of points, you could have factional voting. An example is Democrats don't like what Republicans post: So if there are more Dems than Reps, the Reps get downvoted into oblivion. Now if you had your faction, you could see what your faction felt on the subject not influenced by counter factions. It takes some thought in how to do the server code, or you end up with bad server times or glitchy cases where someone changes their faction, and all the votes change.
God spoke to me.
It's perfect.
PlusFive Slashdot reader for Android. Can post comments.
The correct way of building a site like this is to make sure that the moderation system is web3 compliant, which means that every new user has to go through the strict Natalie Portman filter, where she looks at the picture of the user and those, who she finds acceptable are allowed to join. Those, who she finds attractive are allowed special privileges on the site, they can talk about pouring hot grits down her pants, and some, who are especially lucky, she meets in person, they become the moderators. Of-course those of us, with the biggest heads are forced into compulsory sexual relations with her, and the rest of the users must obey them, including the moderators. This will keep the club tight and exclusive and very intelligent, obviously.
You can't handle the truth.
In addition to the moderation continuum, comments could receive visibility proportionate to the sources they cite. Select the text specifying your statement, and click the (hypothetical) “Prove It” button. The user is then guided through some process to provide material from independent parties. The site running the forum could vouch for sources it deems reputable (e.g., Wikipedia, Congressional Budgetary Office). Users can propose additional sources, and moderators could consider adding them. This differs from free-from hyperlinking in that the citations can not (or are less likely to) be trash. By tying visibility with researched material, competition is introduced among the users to provide factual material that supports their opinions.
I've always wanted to see the ability to tag a response (or the article itself) with its rhetorical type and fallacies in particular. Maybe add some extra categories for "citation needed", "supporting evidence", "contrary evidence" that I don't think would exist in a traditional rhetorical set (could be wrong there). Let moderators (probably users) do more than just say "this post sucks". Let them list WHY it sucks. The moderation could be more concretely proven as accurate or inaccurate, and even better, it might train the readers to spot the nonsense not only in responses but elsewhere as well. Of course, that could be bad for newspapers and terrible for advertisers.
What's needed in most news comment forums is human moderators consistently applying well-defined local cultural rules about what's acceptable.
When those rules are made explicit and then enforced, they can become a self-reinforcing part of the culture. Users get educated in the process, and educate the newcomers, requiring less professional community management.
You cannot replace this kernel of human etiquette with a technological solution and expect to get better "discussions" than, say, here on Slashdot.
So first find the sites that do rules+human moderation well enough already to host the level of civility and discussion you hope for, and distill out the minimum rules and moderator involvement needed to get there. Then add the tech.
Sure, they can be disruptive, but at least they're making an honest, intellectual effort to contribute in their own maladjusted way.
The real problem is your audience. For a general audience newspaper, you get the general public, and unfortunately they're stupid, superstitious, and easily frightened.
The only real solution is to establish a system of unique personal IDs.
Establish a trusted organization that uniquely veries each individual and assigns their IDs. This agency (private or government, but preferably private non-profit) would provide each verified individual with however many unique IDs they need to maintain their privacy. (e.g., slashdot never knows who you are, but they know that your account is unique).
This neatly solves the pervasive problem of corporate entities that use many anonymous accounts and comment-farm trolls. Every account and every comment would be associated with a unique individual.
Voting/karma systems are still useful to weight the integrity of of each individual's posts. Of course the weight of the voting/karma should be based upon the history of whomever is voting.
Totally anonymous posts could still be allowed, but marked/downgraded so they appear later and carry less weight.
Yeah, I know, I am posting this anonymously on slashdot...because I have no idea who owns & manages the slashdot account info database. If we had a good personal ID system, this would not be an issue.
So my undoubtedly unpopular answer is to require verified ID for all posters*. Further require that all posters use their real names as contained in their verified IDs. Even with that you'll still get some amount of trolling/flaming, but it will require much less time to manage than otherwise.
* Before anyone comments on the implicit hypocrisy, you can see that my "homepage" URL contains my name [even if the site is offline]. I don't hide behind an online identity - I am who I am, online and off. Helps keeping me from saying things I'd not want associated with me ;)
Make sure that all kids in public schools get a healthy dose of the dialectical method and reason from first grade all the way through high school. Encourage parents to take courses in language, logic and persuasion at their local community college and to pass that knowledge on to their kids. Wait a couple of generations and we should have a good foundation for sane discourse on whatever comes after the Internet.
What's that? We're encouraging STEM disciplines at the expense of liberal arts education and we need technology solutions to make up for the fact that we're no longer teaching basics of argument that have been around for thousand of years?
Good luck with that.
Slashdot has the issue of anonymous cowards. It has it's benefits, I'm sure, but it makes driveby trolling a hell of a lot easier. Is there any means to report a chronic troll? Is there any backend slash-fu to try to detect sockpuppets beyond IP?
And then you have sites like Boingboing: an idiot mod protecting the fee-fee's of idiot writer's like Xeni and Cory when they make idiot remarks about topics they'd like to pretend they have any expertise on.
Comment threads are effectively a multi-dimensional collaborative filtering problem. I may think a post is "a FUNNY post about CARS", while you may think it's a "SAD post about ENGINEERING". Of course, each of these attributes is not binary -- I may think it's "very about cars" while you may think it's "somewhat sad", etc. When I think a post is "funny" you may be slightly more likely to find it "sad", so if you search for "sad posts", things I think are "funny" should show up.
We have algorithms to solve this (bayesian SVD over sparse tensors), but they're not hugely fast, and not widely known. Nonetheless, this is an implementation problem, not a design one.
And I've never seen anything even vaguely pro-copyright get above a 2.
Right here, in at least two stories there are many counters to your claim. GPL is a form of copyright and people demand that be protected and be upheld.
I think what you are complaining about is that everyone on Slashdot is upset with "The Mickey Mouse Act" and is disgusted that lobbyists determine how long copyright stands so now it's an unreasonable length of time. And yeah, anyone defending that deserves to be modded down. But you're not going to find anybody other than massive studios defending that because why would an artist care that their work is copyrighted past their death? Hell, I would demand it be public domain so that more people could enjoy my work.
You can post positions counter to "group think" but you have to pose them intelligently and try to achieve a neutral point of view when you do it. An example might be proposing copyright reform down to twenty years but enforcing it even more rigorously to ensure that the artist truly gets royalties for those twenty years. Swearing at people and calling them thieves only illustrates you don't understand the nature of copyright infringement nor how the biggest most powerful players have the public by the balls and all politicians in their pockets.
I assure you on copyright and patents, I have often posted comments asking people what they thought a responsible length of time was or asking them how biotech firms should recoup their losses on searching for/developing drugs if they should not be able to patent them.
My work here is dung.
The San Jose Mercury News used to have an awful problem with comment threads erupting into flame wars (as much as I enjoyed getting into the middle of them). That was in the time of anonymous commenting. Nowadays they use the Facebook plugin, so your comment appears beside your profile pic and your name on FB. It has become very civil all of a sudden. You have the option of checking a box to post the comment to your profile, something I never select.
Drill baby drill - on Mars
I did an experiment on Digg and Slashdot a few years back. I posted two extremely similar comments using two different accounts. Then I modded one comment up, and one down.
Keep in mind that there was essentially no difference between these posts other than their initial out-of-the-gate moderations. On both sites, the one I modded down got modded down further. And the one I modded up got modded up further.
While the experiment was quick and informal (I'd love to see others expanding the effort...) it does tend to indicate that as little as one or two people do the "real" moderation of a comment, and many subsequent votes are "me too" votes. Given that, there's not much value in the final tally. And as we see here and elsewhere - unpopular (but factually correct and well written) views can get downmodded into oblivion; or up/down modded so much that the overall moderation is minimal.
Don't allow user moderation like slashdot and digg. If you must allow it, do it very minimally and in a way that doesn't alter whether the comment can be seen by default. DIsplay upmod and downmod totals, and don't assign a final value.
Implement the ideas from this post on k5: http://www.kuro5hin.org/story/2009/3/12/33338/3000
That is the precise problem in all of this. Don't invoke 'public interest' recursively.
"The Constitution, the WHOLE Constitution, and nothing but the CONSTITUTION."
Similar to karma, but much more versatile. As more people meet you and express their opinion of you a clearer picture of you from many different viewpoints emerges. Couple this with tools at the user's disposal to filter or highlight at their own discretion and you have a very dynamic and robust system on your hands which also happens to be reasonably simple to implement.
If Knight-Ridder and Mozilla want the best minds on this, HIRE THEM. They're easy to find. Just look for novel commenting systems. Like, say, Slashdot. Or Usenet.
If they want to re-invent the wheel, conceptualize the now-archaic public square and work from there.
This is my attempt to preserve comment context:
http://threadit.kaizenrails.com/
I think that instead of a list box at the bottom of the page, you have a slider on the side of the page that annotates the article with that users comments. That way specific arguments/comments can be made right in context with the article. Also this will allow for easier detection/elimination of spam because it won't annotate correctly (at least at first).
This would make reading the comments easier because they are in the context of article and reduce the amount of quoting of the original article.
-Lord Shadow
I don't imagine any kind of moderation system, on a news site, will go very far for a simple reason: people don't stick around long enough to engage in discussions. A couple of people comment but the discussion rarely reaches critical mass. News sites are constantly deluged with content so people will be jumping onto the next story fairly quickly. Slashdot's pace is a bit more deliberate, but even here, if you're not on the ball you'll comment late and your post will just end up buried and unnoticed.
Group think is always going to be a problem. Slashdot has one of the best moderation systems out there, and is definitely one of the most balanced in terms of discussion, but even here, the issue arises, especially during campaign season and with certain touchy topics. Still, short of simply letting people post anything they want I've yet to see a better system.
One absolutely crucial element, however, to ensuring the success of the discussion forum is design. Most sites treat commenting as a secondary element. This is not conducive to maintaining a discussion. Slashdot has one of the best approaches because the comments are front and center, and more prominent than the story. It's more of a forum format and each post is very distinct. You don't have them all blurring together like happens on other sites.
Of course, the common issue is that people read only the summary. But the thing is, if you want people to comment you've got to get them into the action right from the start. A compromise may be to make the story collapsible. Arrive at the article and you get the first paragraph, followed by comments. Click on the "more" button to get the rest of the story which expands the whole thing and simply pushes comments down. This way you're not redirected elsewhere for the story.
Break opinion posts down with non judgmental (yeah, right) category moderation. For politics: "conservative", "liberal", "libertarian", "Democrat", "Republican", "SDLP", "Sein Fein" ... etc ... Moderate comments "opinion" - "background" "(attempted) insight" - "interesting but off topic" "funny" "political flame bait" "random flame bait" "incoherent"
Combine this with Facebook login - that alone keeps me off the tech sites that use it - and - here's the important part - PAID MODERATORS who are monitored by karma (or whatever the site is calling it) holding users. Yeah yeah I know web 2.0 is all about crowd sourcing free content but I offer up youtube comment threads as an argument that it's time to move on. Let's get some new buzzwords going and start paying moderators as there are a lot of people out there who could use the telecommuting $10 an hour.
Physics is like sex: sure, it may give some practical results, but that's not why we do it.
inline citations.. linking all the news (and education resources together)...
would be nice for those consipricy theories too.
And yes, my sister is related to the Queen.
and the dollar says 'new order of the ages' which in biblical terms would be 'new order of creation'
Though pyramid saving schemes aren't allowed, someone forgot to tell the bankers though.
thank God the internet isn't a human right.
you will have trolls.
No two ways about it.
You will be supporting "free speech", and people will take it too far, because, frankly, people are immature assholes when they believe they can't be caught.
While allow moderating on comments is useful, it is far from ideal. Too strict of a moderator and nothing worthwhile gets through. Too lenient, then so much garbage gets through that it is next to useless. While I would maintain moderation of comments, I would propose the following system.
First, no anonymous comments. If you are wanting to enter into an actual discussion, then who is discussing should be identifiable.
Second, to entered into a discussion should require one to be registered on the site hosting the discussion. This is related to the first point. Some boards allow you to type in a nickname and email address each time you want to join a discussion. That is basically the same as allowing anonymous comments. Registration may not prove who you are, but at least ties your comments to an identifiable, albeit unknown, person. For protection of the individual, however, the registration system should allow them to keep their data private (ie email address). As such, once a real person has been registered and verified (simple email message/reply), then their online persona exists. You may never know who DcnJoe60 is (my id on /.), but you should know that whoever I am, I always use the same id.
Third, and finally, current moderation systems are simply informational. What is really needed is the ability to blacklist comments from people you don't want to hear from. If you think that DcnJoe60 is just a big blowhard, you should be able to tell the "system" to ignore or not show comments by DcnJoe60. If, on the otherhand, you really like the ideas that DcnJoe60 throws out there, you should be able to tell the "system" to bubble up those comments.
With those three steps, people have the right and ability to say pretty much whatever they want (freedom of speech), but the reader also has the right and ability to only see the comments they choose to. In such a system then, the moderators only need to look at big picture items such as acceptable use (no spam, no vulgarity, or whatever).
This is how things have been discussed for centuries. If we are at the corner pub, the fact that we see each other, even if we do not know each other's real name, removes anonymity. Each of us is free to say what is on our mind. Each of us is also free to enter into or leave the conversation. The barkeep acts as the moderator, keeping fights from breaking out, and other such tasks.
Yahoo groups comes close to this model if you receive individual emails. You may or may not know who a specific individual really is, but you can tie their posts together (step 1 - no anonymity). Step 2 is straight forward, people are registered, at least on most groups. Step 3 is also. For individuals who always post stuff that I don't want to hear/read, there is the delete key (if occasional posters) or create a rule to filter them out.
Yahoo groups does this through downloads. However, the topic is really about online forums. There are no real obstactles, technology wise, from doing this with an online system. The hardest would be the blocking feature, but even Facebook allows for that. What is really needed is taking the already available pieces from wherever they have been implimented and put them together for a complete solution.
preferences? this would be a user abstraction layer with initial comments tagged as something specific that can be recombined by the user to build a unique axis, borrowing from a previous post, that will upmod and downmod. So, comments never have a score until a profile is applied to the tag filter and the comment becomes a bucket for all tags/mods. This will enable profile building(inherent or explicit) for each user this is used to pull out the comments that suit their perspective best. A specific implementation of this would be to use social networking concepts to build a friend tier. A user can see comments through the lens of their collective friend network. It could be thought of as the article/comment genome project - sort of like Pandora. News articles could be sent using this same mechanism with the lens determining articles of interest.
I'd just create a GUI Interface using Visual Basic; See if I can track the commenters' IP address.
because media companies are not interested in the truth, they are only interested in selling copy.
So why would organisations which do not care about the truth want to use a high quality commenting system that would expose their deceptions.
It seems you really need a way to parse out the ideas people are trying to convey and treat those like threads. I'm thinking of my local news site's comment board where most of the posts say the same thing, and also where comments often contain bits of insight and idiocy. You need a way to take a comment and break it up, allowing each point to be debated until it is thoroughly discussed.
Not sure how you'd do that now, but I think in 25 years it's how it will be done.
For starters, how about allowing people to highlight sections of text and rate them up or down - possibly influencing the color or font size(might get unreadable?). That way I don't have to 'like' an entire post - just a few concepts or statements within it.
I'd stay away from reputation-based approaches. Sometimes the crazy trolls have ideas that should be heard.
http://www.masturbateforpeace.com/
Add more than 5 maximum score points and give more points to moderators with better karma. Allow users with good karma to assign more than just 1 moderator point to a particular post and allow moderation to be done at any time (not randomly or at timed intervals). Allow negative moderation in the same way and demote karma based on negative moderation. Detect users who frequently up vote each other to avoid a situation where a small user base could promote each other into good karma. Give diminishing returns from up voting the same users.
Read my short stories - You won't regret it.
Removing or somehow downgrading or burying any posts other than spam for commercial purposes is CENSORSHIP and should not be allowed!
If news organizations want to attract readers (assuming that translates to "make money") they need to do two things - 1) report news that people feel is interesting or relevant. and 2) provide intelligent analysis of the news (not discussion forums).
Too expensive. Plus, it empowers reporters and analysts. This is 2011. Don't you know that investing in humans is a sucker's game?
Comments are like Reality TV. Its much more cost-effective to let attention-starved nobodies provide content free of charge. Plus, it guarantees one train-wreck after another, which drives up ad impressions.
"Broken" comments and forums are a feature, not a bug.
Let's exercise IoC in regards to content!
http://www.alistapart.com/articles/orbital-content/
Instead of users just consuming content, users create content (the comments that news organizations want to farm) and providers consume this content, which is attributed to the user. Users consumer other users' content via the providers, and produce more content in response. At that point providers don't even have to be accurate, they just have to consume the most popular content creators.
Which is exactly how journalism has worked since we got rid of (i.e., started ignoring) that annoying little FAIR Act. Who cares if it's accurate? It sells ads! It's what the White House wants you to hear! It's not like we're working with indelible ink -- the internet might as well be written in pencil, the way we go back and change stuff many times a day, a la 1984. Hell, if it weren't for the comments, I might not even know that the news story I'm reading was changed 5 times before I read it once.
Then again it doesn't matter because the ads will have me forget anything ever happened before the next interruption in my distraction-driven life (soon to be repeated and forgotten). Thank god that other article says Twitter doesn't provide a substantial source of links to news sites. That's exactly what it will be like when we start treating user comments like they're relevant to the news stories.
insensitive clod overlords obligatory xkcd car analogy russian reversals whoosh pedant fanbois ftfy in 3...2...1..PROFIT
There's one major flaw I've noticed in the /. system: groupthink and herd mentality. Anything that perfectly fits a certain mentality will get upmodded, most things that disagree with it get downmodded. Thus, people who disagree with that thinking (or even just don't care about it) have a disincentive to post, and the site attracts people fanatical about that viewpoint, perpetuating the problem.
I agree with you about Slashdot with one exception: a distinguishing quality of the Slashdot metamind is that it's at least subconsciously aware of its own biases. Witness, for example, the fact that your post -- ostensibly critical of Slashdot's readership and moderators -- was modded up to +5. It's precisely that honest admission that the system isn't perfect, and that yes, Slashdot has its systemic biases, that keeps me satisfied with the moderation system as it stands.
When I was following the Fukushima disaster unfolding, I was following 3 sites at once: liveblog at reuters.com, bbc.co.uk, and another. It would be nice if these different resources were merged to provide the most up-to-date information. For example, dynamically linking different resources to make one master resource.
I'm a dyed in the wool socialist, and my angry rants tend to get modded down too (don-mind, it won't stop me :D ). It really doesn't matter what you're expressing, strong political views on /. tend to get modded down. It makes sense, it's a tech site after all.
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
Julian Assange isn't a US citizen. He can break US law all day long and it's A-OK. Now, if he broke INTERNATIONAL LAW or the laws of his country, you might have something there. But whatever law he broke would have to be covered by one of those inconvenient extradition treaties...
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
Force people to only be able to start threads with pictures attached.
Google Wave is the future of Web Communications!
The problem with the /. system is that it's one-size-fits-all. A better system would expose the raw moderation data to the client.
For example, consider a post that is highly informed, reasoned and articulated but expresses some viewpoint which herd sensibilities oppose: So it gets 10 positive and 10 negative mods, and the current system therefore buries it amongst the least significant posts (the system instead elevates a swathe of highly similar posts expressing the majority viewpoint, ultimately just reinforcing one potentially narrow mindset). Unfortunately, the posts I most want to read are the ones that expose me to viewpoints outside of my own. I want to preferentially read the posts that generated a strong-yet-controversial reaction (e.g., ranking posts more according to sum of absoluted mods, rather than the sum of signed mods).
But this is only one alternative ranking scheme. For example, I also want to cancel out the tendency to arbitrarily favour the earliest posts over ones contributed a couple hours later, so I might rank according to rate-of-moderation (calculated using timestamp from the post) instead of accumulated total. Or somebody might develop a Last.FM (librefm) style model, which analyses which other mods tend to agree with the mods who favour your favourite posts, and therefore recommends the posts each of us will personally most enjoy seeing. Or, some people might prefer to build their own trust networks, and preferentially read posts recommended by moderators endorsed by authorities they respect. (For example, on a cutting edge science topic, I'd like to first hear from people who have actually studied related fields, leaving the general sci-fi reference jokes and alternate theory speculations further down the page.)
We have to recognise that /. is already exceptionally good compared to most forums in the world. Try discussing the physics of climate change on physicsforums and you very quickly see the downsides of having some fixed set of "expert" moderators imposed from above (fallibility, autocracy, censorship, etc). The key at /. is automation and democratisation. But it could go further than that, by also democratising the post ranking scheme, and the site administrators only need to facilitate that. (Just like the users themselves perform the labour of eliminating spam from /., so too the users could burden responsibility to evolve superior ranking schemes.)
Sorry, they're busy right now, stealing your identity.
1. Meta-Moderation
> I'd stay away from reputation-based approaches. Sometimes the crazy trolls have ideas that should be heard.
This is why meta-moderation is useful. With no context, a post can be generally identified without a the observer being aware of the specific context. Especially topics the user has not participated in.
2. DONT link users to posts, for readership. While internally you will want to track them, to some extent, allowing people to assemble personas they feel they have to protect and that other people feel they have to oppose, is less desirable than the "community" that can be fostered from it. If your goal is a community built from altruism in sharing, there's nothing to be gained by linking a persona (it's all anonymous) to a history.
Often wrong but never in doubt.
I am Jack9.
Everyone knows me.
At the highest level wold be persons who have registered signing keys with a global scope (e.g. public key infrastructure etc) and used those keys to sign their commentary publicly.
Next are those with same who have signed their posts but have hidden those signatures.
Next are those who have hidden their signature and their user names but still signed.
Next are those who have a user account and a used that account name.
Next are those who have a user account but hidden that account name (AC to the world, non AC to the platform)
Next are those who remain fully anonymous.
This establishes a hierarchy of credibility, int that they have put their identity behind their words. Individual posts from a high credibility source can be, at the time of post, made in a down-mode that appears to be lower credibility.
When an identity is queried it can be cross queried by their real and apparent credibility (key level and down-mode) and in the case of PKI etc, a user or source could back-reference signing keys to the master key as "sometimes one feels like a nut" so there is no reason to ban using separate signing keys for your different personae.
In all cases each posing is made using a computational/web dingus (javascript, java applet, real application, etc) that signs and encrypts the data for transit to the server the fingerprint of the signing key is used as part of the Message ID catalog the message. This way spambots etc can be forced to either create a new fake key for every message or have all their messages invalidated at once when they are caught spamming.
Optionally posts are based on, and must return include-headers that are generated by session.
Optionally services could accept responses based on static forms or just well-formed mime messages with payloads encrypted to published keys.
Response forms and posts can be sent and received via alternative means such as email (allowing for email drop-boxes to be used to defer identity).
The response can include the public key to use to check the message signature to allow for one-time commentary tickets.
While the servers now have the burden of decrypting the messages, the commenter origin is known to have received and processed a request form, and done the exact computational work to generate the message in context (so no bulk messaging by form replay or blind post). Users are encouraged to rise to at least the third level of coherency (key generation costs are best paid once). True anonymity is possible by generating and transmitting unique public keys with each signed message.
The server agrees to decouple and not-record any reconcilable IP or email address with any given message decrypt and post. (basically use a queue to decouple POST action from the validation and site content integration phases of the message.)
This forms a balanced covenant of identity between all parties that even allows cross-site identity (via the PKI key etc) to be maintained when/where the user so desires. It discourages spam, it allows users to check the credibility of sources by finding semi-automagically other commentary by the same persons. At the highest level it requires no "account database" as the key/fingerprint is the identity so nicknames are just nicknames for "real participants". Key banishment makes cleaning out spam and abuse pretty easy as ones current comments risk all of ones prior comments under the same key.
E.g. this makes the user put something in the pot if they want to be taken seriously, while still providing one or more levels of anonymity.
As an added bonus anyone should be allowed to pick up the private replies and messages to any key identity. Such replies would be posted to the account encrypted with the target's PKI key by the sender. This does turn every web site into a private message drop box, but who cares. It does take up some storage on the site but if every site is doing this stuff then no one site would be overburdened. The idea here is to remove the entire
Innocent people shouldn't be forced to pay for inferior software development.
--"Code Complete" Microsoft Press
- only allow real names, this alone keeps a huge portion of discussions on the up and up
- social and merchant site integration, sites where users can be verified in some way, facebook or amazon via a credit card increases a users default good karma
- no displayed point system, or post tally, or userid snobs
- threading is essential, so the replies are tied to the original post
- keep it simple. plain text. uses less bandwidth
- a few questions asking detailed information about the article which allows the user to post comments, what is the point of them commenting on the article if they haven't read it? in any case it will reduce trolling
- users are allowed to permanently hide or upvote other specific users
- comments are shown based on a score calculated based on: user score threshold setting, user verification results, recently posted comments score, number of recent votes to other users comments (the more they read and interact with other users the more they should be seen by others)
- all users can vote on others comments after a certain number of comments are posted, verification is done, or time has passed
some ideas copied from previous comments
consider a post that is highly informed, reasoned and articulated but expresses some viewpoint which herd sensibilities oppose: So it gets 10 positive and 10 negative mods
So you want both a score like the current score and a separate "amount of controversy". Sites using the like/dislike system of Everything 2 and (lately) YouTube show the total number of positives and the total number of negatives. Slashdot used to have something like this, showing how many points of each reason were applied, until an infamous comment in the story Oracle Breakable After All drew over 800 mod points. Then the administrators changed it to percentages. But users managed to figure out controversy levels by the exact percentages (e.g. +3, 67% positive and 33% negative means 9 total points used), so they changed it to percentages rounded to the nearest ten percent. It appears to be the intent of the administrators to hide a posts amount of controversy from users.
So I find a comment that expresses a common misconception, and I wish to provide my view on this misconception. But in order to reply, under your proposal, I first have to read the article. How do you recommend that I read the article if the web server or underlying database server is not responding? And for stories that link to an article behind a scholarly journal's paywall, do you recommend requiring that all people who wish to comment pay $30 for access to a single article?
well-defined local cultural rules
Some of these "local cultural rules" may differ greatly from one part of the world to another. So good luck enforcing them on anything bigger than the web site of a small to mid-size town's local newspaper.
As I discuss here: http://groups.google.com/group/openmanufacturing/msg/576771df555e729f?hl=en :-) ..."
"You may well be right in practice with email, but none-the-less I think this once again suggests the need for much better tools for communications than plain email. If we were using something like a social semantic desktop (or whatever) to create these open manufacturing systems, people could post whatever they wanted to a cloud of ideas and semantic tagging done by the community (including how hardware oriented some post was) would let items that Adrian was interested bubble up to some semantic-related queue he set up for himself, essentially as a form of topical moderation (but, where there was an infinite number of topics). At least, that's the hope.
That was in the context of email, but much the same applies to almost any web setting.
A 21st century issue: the irony of technologies of abundance in the hands of those still thinking in terms of scarcity.
Something like Craiglist's discussion technique could be tweaked. Instead of having 2 main panels, have 3 main panels: one for the full thread list, one for the sub-thread you just clicked on, and the bottom right for the full text.
The far left panel would be the main thread list. The middle right panel the sub-thread list, and the bottom right is text. The slim top would remain the "menu path".
Both panels would be populated when a left message is clicked on.
Table-ized A.I.
....is a great example of how not to do it.
I see even classic Slashdot is now pretty much unusable on dial up anymore.
Reddit has a system to base comment display and positioning on estimated asymptotic upvote rate, rather than raw scores or raw quotients. Randall Munroe from xkcd outlines the thinking here. This enables comments that are late to the party (like this one :) to gain traction and get displayed more prominently, if warranted by vote/moderation patterns.
Such quality estimators might be less important in a system with restrictive moderation caps a la Slashdot, but it illustrates a good point to consider in designing moderation systems.
Another thing Reddit does right is that "best" is not the only way to sort comments. Those who enjoy debate can choose to browse "controversial" posts instead, for instance.
In effect, Reddit recognizes that different readers have different preferences, and provides mechanisms for browsing and interaction that can accommodate these differences. This has the potential to reduce groupthink, or at least let several groups coexist and exchange information on a single site. Catering better to each individual may thus give a better site for all.
Paddle faster, I hear banjos
Does nobody else see the irony of a comment like this being moderated to +4?
The fact that it's been validated by the system it critiques invalidates it.
Only for the intellectually lazy, which flourish on Slashdot, like it or not. Your own post is proof enough of that, but I've enough dealings with moderation-abuse to de-list Slashdot from my bookmarks for this very reason.
http://www.debunkingskeptics.com/
I know this is a slightly different angle from what others have been talking about, but I personally think that things like debategraph.org could be a really interesting alternative to a linear or threaded comment system. This graph-like system has a slightly steeper learning curve (but potentially this could actually help with the problem of trolls). The whole system is set up to let you argue particular points and cross-referencing between comments (including those from different stories) is all part of the process. I also feel that it encourages you to stay longer on the site as you explore the emerging graph.
I've thought about this issue a lot, since becoming a co-sysop in 1985 and for the next ten years, moderating dial-up BBS communities, then a series of web-based communities ever since. I founded or co-founded three popular BBS systems, got to know the community thoroughly, and have about five thousand people (daily) visit the web-based communities I moderate.
In my experience, the solution is an unpopular one. We like invisible hand, quasi-anarchistic, external solutions. Make a rule, and let it do the hard work.
From what I have seen, this is the wrong direction. We don't need more rules. We need organic leadership that can tell the difference between a valuable minority opinion, and a troll or information saboteur.
Allowing the community to police itself means that a minority of the users who spend a majority of their time online, predominate. This group, which I call "loudmouths," tend to shout down opinions that vary with official community hive mind dogma.
Slashdot's system works OK if people know how to game it, e.g. making their posts spicy enough to get noticed and vague enough to not hit the hot buttons of the fanatics. But nothing substitutes for a good moderation team.
If you're really good, you'll have the smartest members of your community at the very top, making the difficult decisions ("is this a future Einstein, or a clever troll?"). They will then have a support network of people who do the more basic filtering for spam, stupidity, etc. This group will be both vertical in hierarchy, and lateral, in that they support each other and fill in for each other when they get busy.
Their role is not simply to prune out the bad, but to uplift the good and to make connections between ongoing threads. They are content-creators that set the tone for the forum by keeping the quality level high.
This, in my experience, is the only way to have a stable and thriving community.
Futurist Traditionalism
After 2 decades of lurking the net, leaving breadcrumbs in forums, irc, nntp, and blogs; I have come to conclude, with others, that the meat of any story lies in the commentary more than the story itself.
More important, the best parts of the commentary may have no bearing on the subject/topic of the story but are often tangential. This is mostly news stories AOT how2's and responses as answers
(like wrongdiagnosis.com or earthclinic.com)
So, sticking to 'news' stories, a story, any comment - my take is that they, if not all writing, are just 'wordballs'; balls of words.
In and of themself, meaningless; the o/p of someone's mind; like trees falling in a forest. They only generates value respective to some 'being there', the 'conversations' they inspire.
But the interesting things is that these stories often start conversations unrelated to the subject;
but far more interesting.
The daily headlines really do not show where the best conversations are. And I don't have all day to spend on /. poring over all news du jour (I've easily spent a couple hours just on 1 story).
what's interesting is that there is no accounting for 'taste'. Nobody knows what someone is interested in. And, what is interesting may lie deeper in a thread than is apparent. (After all, filtering by interest/informed/ (tag) is not always really accurate but only recourse to wading thru perhaps 100's of posts, dozens of pages. Just to find some juicy points; or for me, more links that I would not have thought to find on my own, but at the cost of sending me into another spidey-hole.
So, I've found a way for the wordballs generated by a story to be measured independent of internal workings; the 'rapport' of a conversation, using moderation points, tagging, link refs, etc as sparklines to help the viewer (me) decide where the more interesting conversations are.
This 'meta-system' would present me with a visual
graphic of the days stories measured by the scope and quality of the responses, regardless of actual content.
Ive fleshed out the details of this system and it feels like a fun project. A meta-site (SlashMash)
that graphically overlays a visual map of each headline's activity.
In deference to those who extoll usenet, this proj could even run a Dnews site that lets users meta-moderate this abstraction (based on the wordball URL) which w/could facilitate fact-checking or another part of the system that provides better credentialism; meaning that it knows who is authoritative.
To go further into the deep, combine /. with a slew of other websites, find ways to connect them, and then go hog wild. Or reference wordballs to archived history and, or go 6 degrees from Kevin Bacon. Lots of possibilitites, very entertaining in and of itself as well as a vehicle to finding the gems of the net.
resist propaganda
I think the main issue is the whole linear thread structure which has been used since what....20,000BC?
Ultimately it may be time to move into a 2D, or even 3D comment space. Instead of giving 'mod points', to move a comment up/down a thread, a person would have a 'mod vector' to move the comment around a 3D space. Basic guts of /., but in 3D. Perhaps the space could be broken into liberal vs conservative in one direction, information/solutions vs babbling/trolling in another, and Innovative vs conventional in the third...of course special sections for 'Bush's fault', 'Obama's fault' and any post mentioning Hitler / Nazis. All comments, even trolls, religious fanatics and opposite view points would be preserved, but the user could navigate the space to see comments whether they want a good laugh at trolls, they want to see how long it took for Hitler to be mentioned, they want to view similar trains of thought, or even [gasp] opposing view points.
How is karma relevant in any way when it is maxed out so fast
The alleged misuse of Insightful began after people observed fights between one group of moderators who thought a comment was Funny and another who thought it was Overrated and realized the imbalance. Five Insightful + Overrated have a net effect of 0 karma, but five Funny + Overrated have a net effect of -5 karma. I have read of incidents where an especially controversial funny comment takes a user from Excellent (able to post with bonus) all the way down to Terrible (posts start at -1).
and really, doesn't mean anything?
Moderators are supposed to be chosen from among users with "good" karma.
About trolls: the norwegian TV channel nrk has a webpage called nrk beta (nrkbeta.no). They write about gadgets, cameras, IT, and generally News for Nerds. For some unfathomable reason, I have never seen a single troll post in any of their discussions. I would love to know why, but I suppose I never will.
XKCD #810 already solved this problem. Get with the program Mozilla.
It is interesting to note that, as with most online comment threads, the conversation here *barely* touches the author's original question, instead people posted 500 messages about trolls and moderation :( ...
I've had an idea floating in the back of my mind for a while that would allow each user of a discussion system to see either their own preferred form of groupthink, or for those who want to see lively debate to see a mixed variety of intelligent opinions.
Every user can like (+1) or dislike (-1) a comment. When they do so, the database records that that particular user likes that particular comment; so you have a table of "ratings" which pairs a user to a post with signed bit. Each comment's base score is calculated based on the sum of likes and dislikes. Every logged-in post begins at +1 because it's assumed that the poster likes his own comments; anonymous users cannot cast ratings so their postings start at 0. Allow users to change their ratings of postings, and logged-in users can choose to rate their own posts 0 for the equivalent of "no karma bonus" (because of user-weighted multipliers, see below), or even switch it to "-1" if they decide that they later regret their own comment (since, like Slashdot, there is to be no comment deletion).
Then, every user in turn has a relationship score with each other user, based on how many of each other's comments they have respectively liked and disliked. Relationship scores are multiplicative, calculated as the ratio of likes to dislikes. So say, for instance, that Alice likes Bob's first post; Bob's post gets a +1 to its base score, and Bob gets a +1 to the numerator of his relationship with Alice, making his relationship 2x instead of (the default) 1x. What that means is that thenceforth, all of Bob's posts automatically appear more highly rated to Alice; in Slashdot terms, we might say that Bob has good karma, relative to Alice at least.
What's more, all of Bob's ratings also get this multiplication. So say that Bob likes Charles's comment, and rates it +1; that only adds 1 to Charles's comment's base score, but Alice sees it as +2 because her relationship with Bob is 2x. But that has no effect on Alice's relationship with Charles, which is still 1x. If Alice reads Charles's comment anyway and likes it, in disagreement with Bob, then Charles's relationship to Alice becomes 2x as well; Charles's comment now appears as +2 (presuming Charles likes his own comment) -2 (because Bob dislikes it) = 0. If Alice likes more of Charles's than Bob's, then whenever Charles and Bob disagree on the quality of a post she will see it rated more toward Charles's opinion; or vice versa if she likes more of Bob's comments than Charles's.
The potential to get an echochamber groupthink effect going on here is obvious: if Alice is liberal and likes Bob because Bob is liberal and Bob only likes posts with a liberal bias and all in all the only thing Alice likes, and the things they like, etc, are liberal, then Alice will see liberal comments all hugely modded up. If Charles is a conservative, and Alice and all her liberal friends and all their liberal friends dislike Charles's posts, then Alice will see Charles modded into oblivion.
However, on the other side, if Charles the conservative is disliking all of the liberals' posts, as are all the conservative friends whose posts he likes, and so on and so forth, then Charles will see Alice and Bob and all their liberal cohort modded into oblivion, and all of the conservative posts float to the top. You end up with two simultaneous coexisting groupthinks in the same discussion board.
Where it gets interesting is if you have someone who likes a good lively debate and rates up insightful, interesting, or informative comments regardless of simple agreement/disagreement. Say Dan does this; Dan will then see things which people who he found insightful liked as more highly rated, on both sides of the debate. If Dan thinks that Alice has good points and Charles also has good points, he will see them, and everyone that they like, more highly rated, minus those rabid biased trolls on either side which he has in turn disliked.
Where it gets really interesting is: what if Dan's posts are hig
-Forrest Cameranesi, Geek of all Trades
"I am Sam. Sam I am. I do not like trolls, flames, or spam."
make the whole concept of the website intellectual, and hopefully attract like-minded users.
for example, a site that compares different versions of the same story would be interesting. users would post two or three politically-diverse versions of the same story. both the news and the media itself would be up for analysis.
You mean something like SideWiki provided by Google Toolbar? Features like this have been around for ages, I can distinctly remember installing a toolbar that did something simmilar in 1999. The problem is - you really don't want to know the thoughts of people who have only this in common with you, that they visited the same page. Unfortunately, most of the people who use the internet are either illiterate, stupid or both. Example: comments on YouTube, if you find one that is useful or at least vaguely coherent, please let me know.
What if you got a +1 comment bonus if you make a comment after reading the article instead of not being able to post a comment at all.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
A browser plugin that allows you to sign comments to any site with your PGP key.
A competitive ecosystem of third-party sites developing PageRank style algorithms to categorize signer's quality and relevance in different subjects.
A browser plugin that hides all unsigned comments, and comments ranked below a threshold.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.17 (GNU/Linux)
iEYEARECAAYFAk3KruUACgkQV/+EQ9Wqf+YLlgCfbty8YrbMmCa5JQsSowCyxpCM
ru4An19+8/cSMtPicCyvAiP5M++W8Tq8
=pZCQ
-----END PGP SIGNATURE-----