How To Improve Online Comments: Test Whether People Have Read The Article Before ...

[Mike]

For a while now, Techdirt has been writing about the decision by some sites to stop allowing readers to make comments on articles. We've pointed out that's pretty regrettable, especially when it's couched in insulting terms of "valuing conversations" or building "better relationships." Dropping comments is a lazy response to a real and challenging problem: how to encourage readers to engage in meaningful ways.



As well as a natural tendency for people to write hurtful or insultings things that they probably wouldn't say to each other face-to-face, there's another problem: the rise of Internet troll factories whose entire purpose is to flood sites with propaganda in the form of comments that espouse a particular viewpoint. As we noted recently, Google is looking to use machine learning technology to help identify and then deal with toxic comments:

a publisher could flag comments for its own moderators to review and decide whether to include them in a conversation. Or a publisher could provide tools to help their community understand the impact of what they are writing -- by, for example, letting the commenter see the potential toxicity of their comment as they write it. Publishers could even just allow readers to sort comments by toxicity themselves, making it easier to find great discussions hidden under toxic ones.

As Google itself admits, the issue is "about more than just improving comments. We hope we can help improve conversations online." A rather clever way to do that has been devised by NRKbeta, the technology site of the Norwegian government-owned radio and television public broadcasting company, NRK. Here's the basic idea (via Google Translate):

a small [on-screen] module is presented to you as a reader with three questions from the article that you must answer in order to be able to contribute to the discussion.

Actually reading the article before you comment on it -- pretty revolutionary, no? NRKbeta realizes that it's not a perfect solution:

We know of course that it is possible to "cheat" with these questions by searching the text above [the on-screen module], and that using this approach it cannot be guaranteed that everyone actually read the article, but we still think it's worth the experiment.

It's hard not to agree, because it tries to tackle one of the root causes of comments that add nothing to the conversation -- a failure to read what the article said -- by making it a pre-requisite before you can add your own thoughts. It also has the virtue of being extensible in various ways. For example, there could be more than three questions in the pop-up box, and your comment's place and prominence in the conversation could be determined by how many you get right. This might allow the thoughts of more engaged readers to bubble naturally to the top of the conversation. The fact that the code for the feature has been released as free software makes experimentation even easier. NRKbeta's idea certainly seems a better approach than simply giving up and removing comments altogether.



Follow me @glynmoody on Twitter or identi.ca, and +glynmoody on Google+


Permalink | Comments | Email This Story



Originally Published: Fri, 10 Mar 2017 19:39:00 PST
source

[Sid]

hmmm, clever

[joew771]

I think that most people actually do read the article first. It may stop some spam comments, but it certainly won't stop stupid ones and most certainly won't improve them.

[ID10TError]

I think that comments are self regulating and truly toxic comments will be beaten down by other commentators that disagree, what is really needed is a way of showing popular comments vs showing an unbiased overview of the good vs the bad. How do you rank comments based on popularity without having a like button?

Comment sorting, some comments relate or are the same as another and should be grouped together? I dunno, i didn't fully read the OP and thus maybe my contribution is not fully thought through.

As a second thought maybe its a viewpoint choice and commentators should flag their comments as positive or negative and this will allow the user to see a glass half full kind of world if they wish or maybe your a glass half empty kind of person?

[joew771]

Any sort of rating system for comments is awful. Go on imgur, for example, where people can rate comments. It doesn't show unpopular ones at all unless you ask it to, and they are some of the best ones. Imgur sucks balls. Lot's of sites do that now though and it's wrong.

Either show comments or don't, stop choosing which are 'good' or 'bad or 'appropriate' or 'inappropriate'. That's censorship plain and simple.

And it's not like they delete comments with swearing or whatever, they just don't show ones that people don't like. That's wrong and not very democratic, which is what they purport to be.

And if Google were to implement some such thing as the OP was talking about, god help us all. It's pretty much the end of free speech. Because who's to say what 'improves' an online conversation? Google? I freaking hope not.

[Headbanger]

I kind of like the default Arstechnica comment layout. Comments are displayed by oldest first by default, people can vote a comment up or down. negative ones don't get deleted, they just turn to grey text instead of black text, and there is even a "split opinion" label for posts that get both a lot of up and down votes.

[Sid]

The problem with any sort of rating system is that it's open to innocent misuse and malicious abuse.

And any system which re-orders comments destroys the continuity which is the basis of human communication.

Whereas a system where you have merely to prove that you have understood the opening post (whether you agree with it or not, and whether the system is perfect or not) before you're allowed to comment, and which then allows you to comment freely, would go some way towards solving one of the problems with online conversations without any real negative side effects.

Of course it does make it significantly harder for OP's, since they will not only have to post but also to come up with questions, but in some ways that's not such a bad thing--it will effectively limit the use of the technology to situations where OP's do actually have clear ideas of what they're talking about...providing a useful indicator of whether the post is even worth reading in the first place.

[Moe]

[Clubbin2Live]

I don't know about all this talk of regulating / controlling what people or possibly what non-people say.
Sounds dangerously close to some form of censorship.
Yes, I get the idea of trying to curtail bots or the likes thereof from posting
stupid stuff and then publicizing other stupid stuff. I dislike them as much as you do.

Do we really want to decentralize and remove the control of systems from humans?
I suppose it depends on the system type and how well it's currently functioning and at what cost.

Constructive and helpful may be tricky to measure / judge and do we always want that anyway.
I skip over what I find useless for me knowing well that another person might find something
worthwhile for them out of what I perceive as nonsense.
Sure it takes more time but I do get to see who and what I do not want to associate with
as well as those I do want to be around.

I think that I want the bots and AI to always be detectable by humans.
Now we border on Hollywood Sci-Fi material which may one day be serious issues for humanity.
I mean take a look at last week's X-Files episode. Not really far fetched at all.

Peace

[dueda]

I think that comments are self regulating and truly toxic comments will be beaten down by other commentators that disagree, what is really needed is a way of showing popular comments vs showing an unbiased overview of the good vs the bad. How do you rank comments based on popularity without having a like button?

Comment sorting, some comments relate or are the same as another and should be grouped together? I dunno, i didn't fully read the OP and thus maybe my contribution is not fully thought through.

As a second thought maybe its a viewpoint choice and commentators should flag their comments as positive or negative and this will allow the user to see a glass half full kind of world if they wish or maybe your a glass half empty kind of person?

The main objective should be save time and just get rid of trolls and complete imbeciles. Of course most corporations won't be saints.
With that in mind, just what is constructive or popular isn't what we should really look for, but a balance.

The problem with any sort of rating system is that it's open to innocent misuse and malicious abuse.

And any system which re-orders comments destroys the continuity which is the basis of human communication.

Whereas a system where you have merely to prove that you have understood the opening post (whether you agree with it or not, and whether the system is perfect or not) before you're allowed to comment, and which then allows you to comment freely, would go some way towards solving one of the problems with online conversations without any real negative side effects.

Of course it does make it significantly harder for OP's, since they will not only have to post but also to come up with questions, but in some ways that's not such a bad thing--it will effectively limit the use of the technology to situations where OP's do actually have clear ideas of what they're talking about...providing a useful indicator of whether the post is even worth reading in the first place.

Yes, Planet Earth being the biggest example. It's impossible to completely avoid.

As texting/twitting shows, we're developing new ways to converse; guess I remember an old short work from Poe...

Agreed.

Not sure; AI can take on most of the work and mods are good for it too. Oh, more work for you guys!