Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
Promoting quality discussion on K5

By Dacta in Meta
Thu Oct 05, 2000 at 06:27:43 AM EST
Tags: Kuro5hin.org (all tags)
Kuro5hin.org

I'm worried about the apparent lack of quality discussion in some articles on K5 lately. This is my proposal of how to deal with this. Is is somewhat radical, but I'd like to hear what people think of it.


This is my first Meta-Article. I've tried to steer clear from submitting Meta stories before because I'm never sure how useful they are. However, I have been thinking a lot about the (potential?) decay of the quality of discussions on K5 (and Slashdot), and if techinical means can be used to overcome that trend. I'm generally not a believer in technical means of improving the quality of discussions. I think it is the intention of the people that matters more, but I agree that technical means may help guide those intentions.

Here's my (somewhat radical, I think) take on how to improve the discussions. Even if it isn't adopted, I hope it makes people think about what K5 is trying to achive

Currently, as I understand the K5 story system this is what happens:

  1. A story is submitted
  2. Voting occurs
  3. The story may move to a section only, the front page or be rejected, depending on votes
  4. After a certian number more stories have been voted to the front page / the section page, this story moves off it and into older stuff.
At step 2, editorial and topical comments may be posted. At (and beyond) step 3, topical comments may be posted.

This is (roughly) the way most news/discussion websites work, and have always worked. The open submission queue is a wonderful innovation, but the flow of stories remains the same.

I'd propose it's time to change this. Recently, many stories have been submitted and accepted on K5. There is nothing wrong with this in itself, but it is my view that it is making the level of discussion of some articles suffer.

My proposal is to introduce a story property called "Activity Rating". Here's how it would works:
The submission queue remains the same, but people can only vote for things to be placed on the section pages. Discussion continues there as normal. The stories that generate the most (or perhaps the highest aggregate|average rating) discussion are promoted to the front page, where discussion continues. The story remains on the front page until a higher rated story appears. The story then returns to the Section page, were it remains until its activity rating is such that it is pushed off there into older stuff. It remains there, unless its activity rating increases enough to move it back to the section page.

Why do I think this would be good?

  • It would cut down on the number of "duplicate topic" stories. Instead of everyone submitting "a discussion of digital music distribution" article, or a "Newbies need to learn the K5 way" article, everyone would see those stories on the front page, because they would remain popular.
  • It would promote in-depth discussion. Currently, I think people are rushing their thoughts (like they do on Slashdot), because they know the story won't be on the front page in a days time. This isn't good - K5 should encourage less hasty posting.
  • It would allow new users to learn from the body of knowledge in a discussion, rather than posting a new article because they thought they were the first person to think of an issue.

There are problems, too:

  • A story with a very high activity rating might stagnate. Therefore, activity ratings should decay with age.
  • We already have the (user configurable) "Hotlist" of discussions. I'm not sure how many people use it, though.
  • "Flamefest" articles might end up having a high activity rating, and flamefest are always difficult to kill. I'm not sure how to deal with this, especially if people rate up postings they agree with, rather than good posting. Perhaps stories with lots of postings that some people vote way up and others way down would lose activity rating?
I'm sure there are other things people will think of.

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Poll
How should we deal with discussions
o They are fine now 34%
o Adopt this proposal 65%

Votes: 61
Results | Other Polls

Related Links
o Slashdot
o Also by Dacta


Display: Sort:
Promoting quality discussion on K5 | 43 comments (35 topical, 8 editorial, 0 hidden)
The level of discussion is not that bad (3.46 / 13) (#7)
by intol on Thu Oct 05, 2000 at 12:35:58 AM EST

>I'm worried about the apparent lack of quality discussion in some articles on K5 lately.

Wow! I recently migrated here from "The Other Site" and was pleasantly surprised to find the level of discussion to be quite high on K5. I guess it may be that I have been spoiled by ďThe Other SiteĒ, where the level of discussion has been has been in the pits for years now.

I agree with you when you say that people are encouraged to rush their posts. I personally donít wait very long to post a reply because I know that the story is likely to be removed from the front page within 24 hours. When the story is removed I know that the chances that the story and posts will be read is lower. However, even though I donít give my posts hours of careful revision, I do give each post some careful thought (or at least I try to =P). Even the most prudent users will post some tripe every once in a while. The user may have had a bad day, or may rush a reply because the post/story contained something which they feel very strongly about. Words are very powerful, and do sometimes encourage people to do irrational things. I think that the number of good posts and poor posts increases as the number of users increases. The more users you have, the greater the chance that a user may have had a bad day or be exhausted from work and will post some tripe. However, I donít really see this as much of a problem because I find the moderation system to be mostly effective. I donít know how K5 users view the moderation system, but it may be that I am spoiled by the moderation system of ďThe Other SiteĒ (which I think is quite broken and actually encourages abuse of the system... but thatís another kettle of fish). On K5, good discussion is still possible. I find that the moderation system works well. Most posts tend to hover around ď3" but really good/thoughtful posts tend to get moderated up. Posts that donít really add anything usually end up around ď2". But I do see the possibility for abuse, e.g. I donít think that users should be able to moderate replies to their posts, since the temptation to moderate down posts which they disagree with is too great.

In conclusion I can say that I like your proposal. It sounds like a good idea IMHO. But please consider that even if you were to force users to think about their posts before posting them, you would still get a reasonable amount of tripe. I just got here and the level of discussion seems great to me! You should check out the other sites sometime. They will lower your expectations and when you return to K5 you will feel much better about the level of discussion! (Hehe. Just kidding =P) Yes, I do use the ďHotlistĒ. I hope I havenít bored you all to death by now!

Cheers.

Re: The level of discussion is not that bad (2.80 / 5) (#10)
by Dacta on Thu Oct 05, 2000 at 01:38:18 AM EST

I don't think kuro5hin is broken - but then, I don't think Slashdot is too bad, either. I just think they are different and I want them to remain so.

I know it the users that make the site - I don't believe in techincal solutions, remember?



[ Parent ]
Re: The level of discussion is not that bad (3.50 / 2) (#29)
by jxqvg on Thu Oct 05, 2000 at 11:56:15 AM EST

I just immigrated myself, and definitely think that the quality is much higher. I still feel rushed, though, too. That's because I have an opinion(sometimes), and I'd like it to be heard if I take the time to write it, which to many geeks like myself is quite a task. Maybe even more importantly, I'd like feedback from others, at least a rating to know that someone actually read what I wrote, even a flame gives me a hint at how other people see my opinion.

I like the idea of keeping a topic alive for as long as possible, too, so it can be well discussed, everyone who cares has a chance to say something and discuss it with someone else, etc., etc. OTOH, if a topic stays alive longer, that means there are more topics to pick from, and information overload mode kicks in pretty quickly, people start skimming posts that are as long as mine and yours. I even catch myself doing that on frontpage articles, just because I don't have the time.

<VAGUE QUALITY=VERY>A lot of things in the article</VAGUE> gave me an impression of a cross between a pure discussion forum a la Ultimate Bulletin Board and a traditional news site. Maybe something along the lines of a hybrid would work? I'm done babbling now.

[sig]
[ Parent ]

Might be worth a try (3.72 / 11) (#11)
by TheDude on Thu Oct 05, 2000 at 02:01:54 AM EST

The stories that generate the most (or perhaps the highest aggregate|average rating) discussion are promoted to the front page, where discussion continues. The story remains on the front page until a higher rated story appears.
That's a pretty good thought, there. Such a system would help the stories generating good discussion get onto the front page, where more people would see them, and thus generate more discussion. Recently, I've been missing stories, because too many got posted too fast. When stories only get a few hours' time on the front page, they really can't generate that much discussion. Sure, people can look at their User Info and see if anyone responded to them...and the Hotlist is there for this specific purpose. Even so, something like an Activity Rating would help keep interesting stories generating discussion, as they should.

AR's should definitely decrease comparative to some time length. That's they only way something like an AR would work well. As for flamefests, I suppose that could be a problem. Eventually, though, people who grow tired of the fest (and there will be many, I'm sure) will stop visiting the article, stop posting on it, and the AR of it will slowly decrease. There will surely still be some fanatics flaming each other, but that will probably not be enough activity for the flamefest not to get pushed off the page.

--
TheDude of Smokedot
Drug Info, Rights, Laws, and Discussion
Visit #smokedot on irc.smokedot.org

Why the Emphasis on Discussion? (2.88 / 9) (#12)
by shevek on Thu Oct 05, 2000 at 02:03:18 AM EST

I'm modding this down because i don't think that every article must carry a trail of discussion behind it. some articles (especially MLPs) are not intended to generate any discussion, simply clicks. i know because i recently tried to post an article that was strictly mindless link propagation and it disappeared in a few hours because herds of people "wanted more discussion."

IMO, discussion or activity are only two metrics that may have some bearing on an article. i certainly don't want them to be applied to automagically promote and demote articles.

as far as the quality of discussions, i find them pretty good. there are few trolls, most of the comments are on-topic, and most have significant thought behind them. i'm very satisfied with the amount as well, since the follow-on chat is something i can only occasionally follow in any depth.
-- Philosophy:Cosmology::Signified:Signifier
Re: Why the Emphasis on Discussion? (3.25 / 4) (#22)
by Anonymous Penguin on Thu Oct 05, 2000 at 10:01:24 AM EST

In theory, no MLP will generate much discussion, so it won't have much effect at all. (Unless someone votes an MLP to the front page, in which case it getting demoted because of its low activity is a GOOD THING!)

[ Parent ]
Activity rating (2.50 / 8) (#13)
by charter on Thu Oct 05, 2000 at 02:38:18 AM EST

Sounds like you're basically proposing that discussion articles be subject to mojo/karma? Interesting idea.

I like the paradigm; the level of discussion dictates the location of an article, while the moderation process is deprecated. Not sure how applicable it is to this site, though, since this would probably be kind of difficult to implement with the Scoop engine.

-- Charter



It's not the quality, it's the time scale (4.45 / 11) (#14)
by spiralx on Thu Oct 05, 2000 at 06:42:44 AM EST

Since k5 has been back up it's gained a lot of new users, many of them from /. and yet I haven't really noticed any serious drop in the individual quality of the posts made here. But the huge increase in the number of new stories being posted (I've moderated about 20 in the last day and a bit) is causing a drop in the available time-scale for discussion.

The fact is that once an article has vanished off of the front page, activity is going to seriously drop, whatever solutions (Hotlists etc.) are implemented.

So I agree with your idea, since it will keep stories with discussion on the front page, whilst getting rid of stories that have been ignored - there are a fair amount of stories on the front page with minimal discussion happening on them. There have been plenty of times when I've spotted an interesting reply to a comment I've made, but because it was several days ago it's not worth replying since it won't be read.

Anyway, a good idea, and definitely one of the better solutions to the "death of k5" people here seem to think is occuring. k5 isn't dying, it's just moving a bit quickly :)


You're doomed, I'm doomed, we're all doomed for ice cream. - Bob Aboey

Well-reasoned proposal (2.50 / 8) (#15)
by Precious Roy on Thu Oct 05, 2000 at 06:44:43 AM EST

I admit I also find the rate of turnover disheartening. I usually only read the site on my off-time at work (6 am - 1:30 pm), and sometimes when I show up in the morning the front page is filled with completely new stories that weren't even in the queue when I left work the day before...

Re: Well-reasoned proposal (3.00 / 3) (#27)
by maestro^ on Thu Oct 05, 2000 at 10:51:08 AM EST

According to the 'voting faq,' "The score a story must reach to be posted is a percentage of total user accounts. I'm still new here so please correct me if im wrong. As soon as a story gets a requisite number of votes it will be posted to the page it got the votes for (front page or section).

It seems to me that there should be some sort of time bound to the number of articles that get posted throughout the day. People argue that the articles making it the the front page are at a rate representitive of what the k5 community feels is appropriate, or that it will stabalize to such a level. I don't think this is the case. For example, if approximately the same set of users the size of the minimum number of K5 members consistantly vote on new stories to be posted, the only limit to the amount of stories that can theoretically be posted is the number of submissions. This is where someone who might not be able to reload K5 constantly throughout the day might miss a significant portion of content.

Wouldnt it be better if there was some sort of time limit that restricts that only so many posts can make it to the front page every <fill in appropriate time interval here>. Just like the way the other site works, stories come on intervals throughout the day. The story with the highest qualifying votes gets posted. This provides a method of metering stories so that everyone has a fair chance of seeing them throughout the day. These variables can be tweaked to achieve a certain number of posts to the front page per hour/per day.

mae
---
radioactive space alien

[ Parent ]
Re: Well-reasoned proposal (none / 0) (#43)
by Precious Roy on Fri Oct 06, 2000 at 09:31:44 AM EST

I think you're on the right track, but I have a different suggestion. Instead of limiting the number of stories that can reach the front page in one day, how about requiring stories to stay in the queue for a certain period of time (say, 18-24 hours) to give a fair number of people time to see them?

[ Parent ]
No need to rush (2.14 / 7) (#16)
by Beorn on Thu Oct 05, 2000 at 06:46:04 AM EST

I say, leave it for now. Hopefully the number of front page posts will self-regulate to a level that is comfortable to most of the readers.

- Beorn

[ Threepwood '01 ]

Discussion (3.12 / 8) (#17)
by BinerDog on Thu Oct 05, 2000 at 07:05:20 AM EST

Oddly enough, I have noticed that the most interesting discussion about an article usually occurs while it is still in the submission queue. The last few that have had little discussion on them, what did occur happened primarily while in the queue. By the time the story is posted to section, or to the front page, there is rarely anything new to add. Part of this is because many stories seem to be very closely split on whether to post or not (ie, 246 for, 232 against, 171 don't care)

Maybe we need to look at the voting system on when stories get posted. I am not sure of a solution, especially as there has been much flak lately about poor story quality. Will think on it.


-- The Entity Formerly Known as Frums (Cuz someone nabbed my name on K5) (I want it back :Ģ)
Re: Discussion (3.00 / 3) (#23)
by Anonymous Penguin on Thu Oct 05, 2000 at 10:01:32 AM EST

The easy way would be to use the submission queue comment posting for what it's for - editorial comments instead of topical. What should happen is editorial comments get posted while the story is in the submission queue, and these get deleted when (if?) the story gets posted. Then come the topical comments.

This is how I understand what should be happening. If this were enforced (ie. no topical comments at all allowed while the story is in the submission queue), the problem you are refering to should stop.

[ Parent ]
I agree and here's more (3.70 / 10) (#18)
by speek on Thu Oct 05, 2000 at 07:41:58 AM EST

The problem is that stories are born, their lives are short, and then they die. In order to make a comment you feel people are likely to see, you need to make it early in the story's lifetime. This promotes aggressive attention-seeking behavior, and a feeling of being rushed.

It also reduces post accountability (cause it'll go away soon anyway), and it reduces the impression of ongoing public discussion.

I like your suggestion, but, let's go further. The problem is, the front page listing of stories is out of my personal control. When I'm in a story, I can sort the posts by rating, oldest first, newest first. Why can't I sort stories? I should be able to sort stories by topic, oldest first, newest first (current default), rating, activity (as you are suggesting), and who knows what else we can come up with (open search would be nice).

To do this, current activity metrics would have to be calculated by the system. Voting on a story's quality would remain forever so stories could gain or lose in quality rating as discussion went on.

If this were done, there would never need to be that feeling of being rushed - the story could always be brought back provided it was still active, or very good, or showed up in a keyword search.

--
al queda is kicking themsleves for not knowing about the levees

Re: I agree and here's more (3.66 / 3) (#25)
by interiot on Thu Oct 05, 2000 at 10:19:42 AM EST

I wish I could rate this as a 10.

You pick your sorting criteria based on how much time you're willing to devote. If you have very little time, you want to pick the articles with the most impact for the time spent reading, and the stories that everybody should read because it could instantly change their whole outlook on things. If you have a bit more time, you're willing to be involved in discussions by posting short comments or moderating down obviously bad comments. If you have more time than that, you're willing to spend some time digesting stories that require a lot of thought, and you're willing to put a lot of thought into a comment or three. If you have a lot of time, you're willing to look at stories that are old, yet still have some life in them because there's a good possibility that some good ideas could come out of them if a lot of thought and exploration were put into it.

[ Parent ]

Re: I agree and here's more (none / 0) (#36)
by speek on Thu Oct 05, 2000 at 06:34:40 PM EST

Expanding more on my post....

We should be able to sort on multiple properties, with threshold values as well. ie sort on most recent with a rating threshold of 50% (ie the most recent stories where 50% of the people who voted, voted it up). We could eliminate the story submission queue altogether this way.

Also, stories should have multiple topics. The author should have the ability to enter multiple keywords to index the story under. Possibly users should have the ability to add to those keywords.

Disadvantages - more complexity in generating your personal homepage. Easy solution - have a bunch of preset sorting and filtering options such as:

  • Newest, threshold of 100, 75% (newest stories that 75% of voters voted up with an absolute score of 100 or better)
  • Just newest (to see and vote on new submissions)
  • classics - sort by rating, and rating threshold of 75%
  • Where it's at - sort by highest activity
  • and so on
Users could choose one of these as their own personal default home page, or they could invent their own sorting criteria.

Damn, I'd be really excited if K5 did this!

--
al queda is kicking themsleves for not knowing about the levees
[ Parent ]

Separate Section (2.16 / 6) (#19)
by revird on Thu Oct 05, 2000 at 09:01:44 AM EST

Why not keep the front page the same, but have an alternative front page (or a new section) that displays articles as proposed?

Re: Separate Section (3.50 / 2) (#20)
by spiralx on Thu Oct 05, 2000 at 09:08:10 AM EST

Because if you're showing different people different things then that will lead to even more fracturing of this site and even less people getting involved in stories. To as big as an extent as possible, you need to present users with a uniform interface...


You're doomed, I'm doomed, we're all doomed for ice cream. - Bob Aboey
[ Parent ]

Uniformity vs. customization of the front page (3.66 / 3) (#24)
by maestro^ on Thu Oct 05, 2000 at 10:15:03 AM EST

i agree that a front page should be the same for all users. speek suggested in this (#18) previous post that the home page should be customizable to allow users to sort by newest first, rating, activity, # of stories to show, etc. this is a great idea... for a user customizable section, but not for the format of the front page. a front page should be what ties everyone together. it should bring you a summary of the major topics currently active (be it chronological or number of comments) on the site. this initial exposure to the same content creates a bond between all the separate subgroups and subinterests within the site that you might be interested in.

let me take the case of yahoo for example. users are allowed to customize a 'my yahoo' home page for themselves. the home page of yahoo is the same for everyone. yahoo users can then go through the 'my yahoo' link to login and have their own personalized content delivered to them. now, they have the option to cheat by specifing that they would like to go directly to my.yahoo.com, but they are opting out of that shared content. now yahoo is a pretty bad example because theres not a lot of commonality between its users like there is here at k5.

the extreme case is that i can set a very narrow preference to what i want to see on my customized page. this is that bad 'fracturing' (#20) that spiralx mentioned. i'd like to see the ability to sort stories, so one comprimise would be to have a separate page that you can easily get to from the front page that would deliver your customized new story settings to you. by making it easy to get to, say with a prominent link once you're logged in, it would encourage people from simply bookmarking the page. no one should be forced to read something they dont want, but we'd like to inhibit splintering of people into smaller groups as much as possible.

mae
---
super hero for hire

[ Parent ]
WikiNow (1.71 / 7) (#21)
by Sunir on Thu Oct 05, 2000 at 09:39:29 AM EST

Gee, you webloggers. You always want more news but want it to last longer and be of high quality.

Try a wiki. Wikis are in a perpetual state of WikiNow. Everything is always available. You don't have to race to comment. Feel free to take days, weeks or even years to come up with a reply.

If you like meta, try MeatballWiki.

"Look! You're free! Go, and be free!" and everyone hated it for that. --r

How to calculate activity rating (3.00 / 6) (#26)
by Anonymous Penguin on Thu Oct 05, 2000 at 10:32:45 AM EST

Hi

Great ideas! IMHO the easiest way to calculate an AR would be (number of comments)/(time since story was posted). By doing that, you wouldnt need to program in something to decay the AR, because it is doing it automatically (as time -> infinity, AR ->0)

Re: How to calculate activity rating (3.00 / 1) (#32)
by maestro^ on Thu Oct 05, 2000 at 12:28:31 PM EST

Allow me to suggest a revision to the calculation. Suppose people are generating comments on a story at a rate of exactly one per hour. If this rate remains constant, I would argue that the Activity Rate is also remaining constant. Using your formula, we would be able to see that indeed the AR is remaining constant. However, lets assume that no one posts any comments for a day and on the second day they begin posting at a rate of 1 comment per hour. I would say that the AR has gone up from 0 to 1/hour. Calculating the rate based on the total comments over the total time in the system, you would not be able to see this. I suggest that taking smaller intervals, say 1 hour or possibly 2 or 4 depending on volume, will give a more accurate view of the current activity on a thread.

AR = # of posts new posts/short unit of time

If you see my other comment you'll see why this need not be a decaying function.

--mae

[ Parent ]
Re: How to calculate activity rating (none / 0) (#42)
by BlckKnght on Fri Oct 06, 2000 at 01:50:39 AM EST

Better yet, do both. :-)

Value= #ofCommentsToday * Total#ofComments / Age
Maybe it should be Age squared....

Also see my post under the thread "A Front Page Problem" where I discussed other possible measures of value. What do you thing of them?

-- 
Error: .signature: No such file or directory


[ Parent ]
What truly concerns me (3.14 / 7) (#28)
by simmons75 on Thu Oct 05, 2000 at 11:04:15 AM EST

A forum like kuro5hin.org is usually ultimately controlled (well, not really...) by the end users. I've been concerned with the number of toplevel stories I've seen that begin "I've been concerned with the lack of quality in [blah]." And frankly, I feel insulted every time. I'm being told I'm too stupid to know I'm doing wrong. Why am I being told that? Because what I feel is right doesn't correlate 100% with what you feel is right.

Leave it as it is. If an issue generates enough interest, it will live longer.
poot!
So there.

Perhaps you are too stupid... (4.00 / 2) (#39)
by krisjohn.net on Thu Oct 05, 2000 at 08:18:57 PM EST

The subject line isn't meant as a personal attack, but instead a summary of the problem. I've recently come to K5 from /. and I must say that I find the discussions here to be, on the whole, exceedingly trivial. I tried another Scoop-based system, Should Exist and found the same problem, plus a huge amount of duplication.

I believe that only strong editorial control will increase the quality of the articles on any system. Scoop relies on the popular vote. Quantum Mechanics isn't popular. Ranting about "That guy that done me wrong" is popular. Without editorial control I find that discussions devolve done to "birthdays and weather". Sure, the majority will become annoyed that their trival stories aren't being posted, but remember: Half the population have below average intelligence.
Chris (Kris) Johnson

If you like this, try my Editorial -- updated Monday and Thursday.
[ Parent ]

Are stories making it to section-only? (4.33 / 6) (#30)
by El Volio on Thu Oct 05, 2000 at 12:01:38 PM EST

I think the voting system could clear this up, if it were more clear how stories go section-only. Looking at K5 now, I can't seem to find any stories that went section-only and never hit the front page. Maybe I've just missed them, but even so, there can't be more than a couple. (Somebody please point it out if I'm wrong!)

Clearing that up (e.g. take away "don't care", since it basically means the same as "+1 Section") would help a great deal.

I'd also like to point out that the latest incarnation of K5 is still new, and needs time to reach an equilibrium. The "K5 died again this week" stories are missing a key factor, that what we're basically doing is positive feedback (in the engineering, not psychological, sense). Because of that, we're essentially trying to adjust a system's behavior before we really know what it is. Straighten out the FP vs. section voting, and let things settle out. The proportion of rants will achieve a better proportion, story quality will improve (since they'll have to either work harder to make it to the FP, or they'll satisfy a smaller niche of users in a section), and All Will Be Right With The World(tm).

Re: Are stories making it to section-only? (3.00 / 1) (#37)
by Dacta on Thu Oct 05, 2000 at 07:06:32 PM EST

Yes, they are making it section only.

A lot of the more interesting stories recently have been section-only. There was a story on Content-creation which was not front page, and I know the only other previous post DoS story I posted was only MLPed (which is what I wanted)



[ Parent ]
A front page solution (4.42 / 7) (#31)
by maestro^ on Thu Oct 05, 2000 at 12:01:58 PM EST

From #28:
If an issue generates enough interest, it will live longer.

Articles on the front page live only as long as there arent any newer articles that replace them. In the present system, there is no protection against a story getting bumped from the main page even though there might still be a lot of activity within it. Thats why I feel there should be some sort of metering of stories that get posted to the front page.

I know of 2 schools of thought that deal with the presentation of 'active' content. One is to simply post stories in choronological order and as new stories are posted, the old ones age. No priority or protection is given to stories that create more discussion than others. In fact, their time on the front page is completely independent of how much interest is generated. It can be replaced by several stories of moderate interest which have the possibility of sitting there for a much longer period if the submission queue is low.

The other thinking is to always post the most recently commented story at the top. This works well for bulletin board sites that are focused on maintaining a continual discussion on one topic. UBB sites operate in this manner. In this schema it is very easy to identify the most popular topics, they all move to the top as the less frequently discussed topics sift down the list. There are problems with this approach. It is not feasible for a site that needs to propagate new information in a timely manner as any post to a very old article can move it back up to the top where it will sit until it is replaced by other currently discussed threads.

What I propose is a hybrid of these two, with a little dash of spice thrown in. I suggest some assignment/reward of time on the front page to the merit/amount of discussion/popularity a story. This would be a weight based on an Activity Rate. If a story's presence on the front page was only a function of AR, we would not be able to insert new stories onto the front page (there would be no rate to calculate if you assume that the submission queue comments dont count) without having to generate some inital bogus AR.

An additional weight would be given to the story representing its age. New stories would start with a high Age weight (giving preference to newer stories), and a low AR. Their total weight would be calculated and stories would rank on the front page from high to low based on this sum. As a story ages, its AR will go up and its Age weight will go down. If the Age decreases at a higher rate than the AR raises, the total score will go down, allowing for another story with higher total weight to take its place. If a story is very popular, its AR will remain high allowing for older, more popular stories to remain on the front page for a longer time. Less popular stories will fall off much sooner as a result of the linear aging function.

I think this goes a long way in satisfying these conditions that we "always want more news but want it to last longer and be of high quality." i hope this wasnt too complicated, im still new to this whole idea that people actually read posts. coming from that other site, i had forgotten what its like before the trolls and the sheer volume keep me from posting anything of value as it almost always gets lost in the noise.

--mae

Re: A front page solution (4.50 / 2) (#34)
by kallisti on Thu Oct 05, 2000 at 02:02:34 PM EST

Right now, the front page is trying to serve two goals, letting people know the latest stories and serving as the main point of discussion. Perhaps the stories with the most activity could be moved to a "Hot" section or something like that. Users could decide whether to default to Current or Hot (I don't like "Hot", but it gets the idea across). Those who aren't logged in would just go to the existing main.

Mixing the two just leads to frustration for people looking for one or the other.

[ Parent ]

Re: A front page solution (none / 0) (#40)
by BlckKnght on Thu Oct 05, 2000 at 10:41:17 PM EST

Wow, I come back from my CS exam tonight and I see a problem that cries out to me to build it an algorithm (or at least a set of formulas).

  • "Hot"ness could easily be calculated by the rate of posts. A story that has 20 comments in one hour is "Hotter" than one with only 10. I don't know if this would be practical to code (are comment rates logged?).
  • "Current"ness is simply age, or maybe (1 / age to make math better). A story in it's first hour (or day or whatever) is (obviously) more current than one that's been up for several days.

There are other qualities that might be worth calculating to rate stories:

  • Quantity. The sheer number of comments.
  • Diversity. The number of unique people who post comments.
  • Quality. This could work several ways. It could be the average rating of the posts, the sum of all the ratings, or something else along that vein.
  • Instantaneous Quality. The quality that of the posts being made now. This could be the sum (or average) of the moderation done in the last hour (or day). This also judges how much moderation is being done on the comments (not much moderation == lower quality).

      Now, these different factors could all be combined by adding or multiplying them together and varying their relative weights. Maybe the weights would depend on the section. Rants would obviously encourage hotter discussions, while Culture might want more diversity and News more currency. The front page could have some nice blend of all of them, or even be user customizable...

      What do you think?

      -- 
      Error: .signature: No such file or directory


      [ Parent ]
Word Count Average to Prevent Flame Wars (2.50 / 4) (#33)
by OKolzig37 on Thu Oct 05, 2000 at 01:47:32 PM EST

"Flamefest" articles might end up having a high activity rating, and flamefest are always difficult to kill. I'm not sure how to deal with this, especially if people rate up postings they agree with, rather than good posting. Perhaps stories with lots of postings that some people vote way up and others way down would lose activity rating?
First of all, great ideas.

Perhaps to solve this problem you could base it on average word count. Flamewars often can degrade into short posts. Maybe you could take an average of the word count in posts from other threads and determine if the word count for the particular thread is sufficient, beyond some made-up threshold.

In other words, if the average thread-word-count for average threads is, say, 250, and a deteriorating into flames thread had about 100, it could discount that.

Again, not perfect, but it could be a start.
Oldy moldy, history mystery!
Re: Word Count Average to Prevent Flame Wars (none / 0) (#35)
by ODiV on Thu Oct 05, 2000 at 03:58:28 PM EST

Another idea for guessing at whether a certain discussion is a flamefest or not would be to use the moderation system in k5 already. If a discussion has a high activity, but they're all 1s, then it's most likely a flamefest. I don't know how much the ratings are used though and I'm not sure if someone's likely to wade into a bunch of flames just in order to give them all 1s, but it's an idea.


--
[ odiv.net ]
[ Parent ]
Re: Word Count Average to Prevent Flame Wars (none / 0) (#41)
by Dacta on Thu Oct 05, 2000 at 11:26:36 PM EST

I think I'd prefer the "lots of posts with large disagreement about their rating" metric.

The problem is that I suspect a lot people are (or will) vote for things they agree with, rather than for things that should be modded up.



[ Parent ]
Thoughts on user customization (3.00 / 1) (#38)
by Dacta on Thu Oct 05, 2000 at 07:16:18 PM EST

A lot of people are suggesting having user-customizable options to decide how stories are shown on K5

I like that idea, but there is a significant problem, which is why I think my proposal should be the default sorting.

The problem I see is that people don't always customize things - either they don't know how, or they want to see the default view so they can see "what everyone else is seeing".

That is why I don't think the "hotlist" is as effective as if might have been.



Promoting quality discussion on K5 | 43 comments (35 topical, 8 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest © 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!