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

Notes Toward a Moderation Economy

By localroger in Meta
Tue Oct 28, 2003 at 10:42:43 PM EST
Tags: Culture (all tags)

Although I'm filing this under Meta it is not a specific suggestion for immediate changes to Scoop. Rather, it is a set of ideas I've been mulling over based on what e-community engines like Scoop, Slashcode, and various web BBS packages are all trying to become, and how the next generation might do it even more effectively.

Whether you call it Mojo, Karma, "Standing," or something else, all content rating feedback systems have some sort of currency. While there are many different ways of acquiring and spending such capital, nobody seems to have implemented an economy varied enough to be robust. And this is the key to building a system which can be stable in the long term.

Speaking very broadly, any web rating system is trying to encourage certain behaviors and discourage others. Behaviors most e-community operators would want to encourage include:

  • Writing original stories and comments that are interesting or useful
  • Leaving "attaboy" or "rubbish" comments that are "correct"
  • Rating stories and comments "correctly"
Note that by "correct" I mean in ways that most readers find agreeable or interesting. Depending on the audience, this might not mean "factually correct." Activities most web operators would discourage are complementary:

  • Crapflooding
  • Trolling
  • Flaming
  • Straying off-topic
  • Moderation "bombing"
  • Gaming moderation to elevate the status of crapfloods, trolls, flames, etc.
One way of guaranteeing the tone of a site is to have actual human volunteers or employees read it and enforce the rules. This is, however, expensive in various ways. The goal of public moderation is to minimize the need for intervention by people with special privileges. The mechanism is a currency system that increases your capital if you are behaving correctly, decreases it if you're not, and provides some privilege which you can spend your capital on that makes it worthwhile to acquire equity.

Let's look at how Slashdot's economy works:

  • Sparing amounts of currency are distributed by lottery and are used up by rating content.
  • Bad behavior noticed by a human moderator can get you canned from the mod point lottery.
  • Well-rated posts are more likely to be read.
  • Joining the site and being well-rated cause your posts to start out more highly rated by default.
In the Slashdot economy the ultimate prize is having your posts read, as there is a real likelihood that a late comment rated only +2 won't be seen by most readers. The Slashcode system can be gamed in several ways which have been Band-Aided, most notably by limiting equity to 50 points. As in real life, some users became so wealthy they could pretty much get away with anything.

Now let's compare Kuro5hin's economy (which we will see wasn't really changed all that much by the recent modifications):

  • Rating privileges are infinite, unless you really piss off a human operator
  • There is really no consequence other than to your ego for having posts rated high or low
  • A high enough score turns on "trusted user mode" which has few discernible privileges
  • A very low score can hide a post, but not to "trusted users" so this is also pretty meaningless
  • There is an entirely separate one account / one vote system for voting on stories
  • Creating an account is free, so there is no barrier at all to creating multiple accounts to give one's self a larger say in either comment moderation or story voting
  • Poorly rated stories are hidden
The main reward for leaving "good" posts on K5 is an abstract number, but there is no real feedback for leaving either good or bad posts. So the system is entirely open to being gamed (e.g. by users like drduck, using a script to rate everything zero) but this isn't such a huge problem because ratings don't mean much, anyway. The most noticeable real problems are the use of multiple accounts to hack the voting queue, and the very weak feedback to discourage trolls and crapflooders.

Moderation on Slashdot tends to be hit-or-miss, since there are a small number of moderators working; it's common for a good but late post to remain buried at +2. On the other hand the Slashcode system is pretty hard to game. K5 ratings are usually more meaningful than Slashdot ratings because there are more potential moderators, but the K5 system is also much easier to abuse. If good ratings were useful for anything this abuse would doubtless be much more widespread.

An economy, like an ecosystem, is more stable if it has multiple feedback pathways. Any single feedback pathway is prone to catastrophe; the cockroaches eat all the bamboo and subsequently starve, or the one company with all the money fires all its own employees to save money, but they're also its customers and as a result the economy goes bankrupt. But if there are many species with interlocking relationships, or many participants in the economy, a catastrophic turn in any single path does not ruin the system.

With this in mind, let's consider some additional reward systems that could be automated in an electronic community:

  • Equity should be worth something, so that one has an incentive to use a single account and keep it in good standing.
  • The example of K5 shows that new accounts should be created with some equity, as the ability to come in and participate immediately attracts users.
  • There should be sanity checks. There is no reason at all to allow an account like drduck to moderate faster than anyone can read.
  • Leaving highly rated posts should grow one's equity, and leaving poorly rated posts should shrink it.
  • Rating itself should cost some equity, so that one thinks before doing it.
  • Extreme rating might cost extra, so that it is meaningful. For example, the old 5-point rating system was eliminated because most people rated 1 or 5; this could be fixed as follows:
    1. This was so awful I was willing to spend 2 points kicking it.
    2. This was pretty bad. Cost me a point to say so.
    3. Read it, nothing special. Costs nothing to say so.
    4. This was pretty good, worth a point to say so.
    5. This was so good I was willing to spend 2 points saying so.
  • A Slashdot style system might be used to make highly rated posts more visible.
  • Equity might also be spent voting on stories. There are two ways of looking at this:
    1. A high equity might make your vote count more
    2. You might decide how much equity to spend kicking or elevating a story
  • It should go without saying that getting a story voted up should be worth a healthy shot of equity, and getting it to FP should be worth even more.
  • Leaving a diary entry should cost some equity. On the other hand, as with comments it should be possible for readers to spend equity rating diaries so that leaving a popular diary would be encouraged. Unlike comment ratings, I would tend to make diary ratings positive-only, so as to offset the fixed cost of using the diary feature.
  • It might be possible to spend equity on site features that are normally associated with paid membership, such as searches or diary watches.
  • Conversely, the site might sell equity outright for dollars under reasonable terms discouraging abuse.
  • In order to "charge up" the economy it might be worthwhile to dispense equity simply for loading pages (at a low, humanly reasonable, capped rate) to encourage the "silent majority" of lurkers to participate.
There are probably other feedback pathways that could be explored, and some experimentation might be needed to work out details like how much equity you should receive or spend for what eventualities, or whether some streams should be limited or capped. But let's see how such a system might work:

  1. You could only crapflood or modbomb until you run out of equity. If you're not doing anything constructive, that ends the fun pretty quickly.
  2. Because new accounts come with a limited amount of equity, there is a limit to the usefulness of creating alternate accounts. This would eliminate the need for IP crosschecking, which can be evaded anyway by using proxies.
  3. Trolling runs your equity out fast, because pissed off people are more likely to spend extra points kicking you. On the other hand if you are a genuinely entertaining troll, you can get along.
  4. Poor equity isn't just a number; you notice it in terms of lost conveniences before you're effectively kicked off the site.
  5. The "rich bastard" problem is minimized because there is a constant trade in equity; even if you acquire a big pile of equity points there's a limit to how much damage you can do with them, and being obnoxious causes you to deplete equity much faster than you can re-acquire it.
As I said before this isn't meant to be a specific recommendation; a lot more thought needs to go into the nuts and bolts and feedback parameters before a system like this can be implemented. But hopefully this will provoke some more thought and discussion and maybe as sites like K5 and Slashdot continue to scale upward we will see some tools emerge that can keep them useful as they do.


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


Favorite Rating
o Mojo 37%
o Karma 13%
o Slack 15%
o Equity 8%
o Standing 4%
o Rating 0%
o Post it / Dump it 20%

Votes: 45
Results | Other Polls

Related Links
o Slashdot
o Scoop
o Kuro5hin
o Also by localroger

Display: Sort:
Notes Toward a Moderation Economy | 155 comments (140 topical, 15 editorial, 1 hidden)
Very interesting ideas (2.20 / 10) (#2)
by leviramsey on Tue Oct 28, 2003 at 03:41:39 PM EST

I would presume that comment posting would also have a cost, as would submitting stories to the queue.

Actually, I think a reasonable rate of comment posting should be free (say, no more than 24 comments per 12 hour period or something) with each comment above that amount being charged.

I'd key it to moderation (2.50 / 6) (#5)
by localroger on Tue Oct 28, 2003 at 03:44:09 PM EST

You don't want to discourage discussion too much, but you do want to discourage crappy discussion. I was thinking that comments would be free by default, but ratings would reward you.

But then again, a fixed negative cost might be useful to keep the bandwidth reasonable, as I figured for diaries, especially as the site scales.

What will people of the future think of us? Will they say, as Roger Williams said of some of the Massachusetts Indians, that we were wolves with the min
[ Parent ]

Ugh. (2.25 / 4) (#38)
by sllort on Tue Oct 28, 2003 at 07:03:51 PM EST

You don't want to discourage discussion too much, but you do want to discourage crappy discussion.

Republicans think Democrats comments are crappy discussion. Democrats think Republicans comments are crappy discussion. Creating a system which allows users to decide what's crappy discussion and what isn't creates stalemate between two balanced warring factions, or as is more often the case, complete repression of one faction by another.

It's just a god damn dumb idea to give "discouragement power" to users.
Warning: On Lawn is a documented liar.
[ Parent ]

You forgot one important type of user (1.75 / 8) (#92)
by partykidd on Wed Oct 29, 2003 at 04:50:38 AM EST

Users like drduck thinks almost everything is a crappy discussion and should be supressed.

"It is the mark of an educated mind to be able to entertain a thought without accepting it." - Aristotle

[ Parent ]

Money (1.37 / 8) (#7)
by A Proud American on Tue Oct 28, 2003 at 03:46:27 PM EST

Bullshit.  If Rusty were ever to start charging people to use his site, he better damn well start paying those of us who contribute to his site.

The weak are killed and eaten...

[ Parent ]
In this plan (2.50 / 6) (#14)
by leviramsey on Tue Oct 28, 2003 at 03:57:51 PM EST

...rusty is effectively paying us to contribute, using the same medium by which we "pay": the K5 private currency.

[ Parent ]
Good Ideas (2.16 / 12) (#3)
by pmc on Tue Oct 28, 2003 at 03:43:09 PM EST

Somewhat similar to a dicussion I had recently about discouraging trolls, but that died (It was RFC: Fixing the Rating System).

Here is the main one again:

Nuke the account - same rules for everybody. Behave reasonably, or take a hike.
For drduck it's even easier - contribute before you critisise. Not stories (I think that's too far), but comments.

Being of a mathematical bent I probably look to do a system where "comment rating outliers" (i.e. people who are rating significantly differently from average) are identified, and dealt with accordingly.

I appreciate Rusty's comments about an arm race, but, let's face it, it's an arms race or capitulation - ignoring the trolls just leaves you swimming in shit. Which was my point - I don't even have rating turned on now, so useless has it become.

And, lets face it, there are lots of fun things you can do with trolls - the more trollish the system rates you (your level of "trojo") the more likely you are to get the d'oh page. After all, this is K5, it's not like anyone will notice. Higher trojo, randomly drop ratings - a proper user has a 99.999% chance of rating successfully, a high trojo user - drduck - has a 1% chance: "Sorry, rating has been randomly refused. You cannot rate for the next 5 minutes." Or don't even bother telling them. Random page delays. Set the cookie lifetime to 20seconds. Drop 10% of the words from their posts - a finly crafted rant turned into gibberish (of course, this may turn some gibberish into a finely crafted rant, but it's a chance I'm willing to take).

There are lots of tactics that can be used if you really wanted to discourage trolls.

This line of tactic has the disadvantage that new people cannot interact fully quickly, which your system avoids.

But I just offer it again to as food for the dicussion.

Let me just say (2.50 / 4) (#28)
by sllort on Tue Oct 28, 2003 at 06:33:15 PM EST

That this comment is as idiotic now as it was then. If you institute agressive "anti-troll" technology like this, the only losers are the legitimate readers. It's an arms race with only one loser: the people who use the site.

Slashdot's attempts at this have been a gigantic, notorious, flaming disaster, the only outcome of which has been to make a system with filters on posting so restrictive that just cutting & pasting text is nearly impossible, where the journal section is filled with user diaries (from normal users) that read "why did I just get IP banned from Slashdot??". And that's with a few years of effort and a multi-man dev team.

Any attempt to measure what you call "trojo" will end up with the people you call "trolls" reading the source and using the system against you. The people facing automated measures that silently break the site will be the legitimate users.

For a site whose userbase seems to be so against war, it amazes me how many people don't understand that conflict only breeds more conflict. Oh well.
Warning: On Lawn is a documented liar.
[ Parent ]

Let's see (2.00 / 4) (#39)
by pmc on Tue Oct 28, 2003 at 07:05:48 PM EST

Your full-on wisdom from that discussion was

"The only way to win is not to play."

Together with "Slashdot tried to ban trolls, and that failed, so let's give up."

Oh - one other thing:

It's an arms race with only one loser: the people who use the site.

If things continue as they are the only people who will use the site are the trolls - I don't mind them losing.

[ Parent ]

As you wish (1.50 / 3) (#41)
by sllort on Tue Oct 28, 2003 at 07:09:58 PM EST

"If things continue as they are the only people who will use the site are the trolls - I don't mind them losing."

If this means you're leaving if the site doesn't change, allow me to show you the door.
Warning: On Lawn is a documented liar.
[ Parent ]

Nice Door (2.16 / 6) (#91)
by pmc on Wed Oct 29, 2003 at 03:30:30 AM EST

I'll be using it in about 70 days, when my subscription runs out. You see, I've already lost. The articles that used to make the site interesting have, by and large, gone. A large number of the posters that made the site interesting have gone.

I've even got a few stories here that I had planned to post, but why bother? Could I go to someone I know and say "Hey, I've posted a couple of articles on K5 - have a read."? Hell, no. Now I'd be embarrassed to have people know I read this. This did not used to be the case. I don't even have the defence of "Well, trolls are a problem everywhere, but they are trying to do something about it.", because they are not.

K5 has turned from a "Cool Secret" to a "Guilty Secret".

I may pop back to watch the carnage - sort of rubbernecking on the information superhighway - but I won't be resubscribing. Comments will probably be restricted to the "I told you so variety", although I reserve the right to mock in other ways.

[ Parent ]

That "fine line" (2.40 / 15) (#4)
by A Proud American on Tue Oct 28, 2003 at 03:43:58 PM EST

Your analysis is nicely done and quite thorough.  However, I feel that you may have isolated just one particular component of a Web site (moderation); the fault here lies in the fact that other components are affected by one another, and moderation cannot successfully be evaluated on its own without some error.

For example, the ability to rate any and all comments and stories and have as many accounts as you want here at K5 may be a very poor design from a strictly Moderation perspective, but from a Popularity perspective, this might tend to encourage more frequent visits to the site by more people.

Someone like Rusty wants his site to be both respectable and fun for his users; if it's not fun, he won't have any users for much longer, and even a site with a perfect, Grade-A, 100% beautiful Moderation component is thereby deemed irrelevant simply because no one visits it anymore.

The weak are killed and eaten...

Very true (2.33 / 6) (#19)
by localroger on Tue Oct 28, 2003 at 04:13:49 PM EST

I think it is very important not to discourage behavior that is desirable. That's why I think you should get some equity just for showing up. But once you're here and you're past that hopefully very low barrier to entry, there should be clear signals encouraging what the site is supposed to be, and discouraging what it isn't.

I think it is a very important feature of what I'm proposing that it doesn't make trolling or crapflooding impossible, it just makes them expensive. So yes, we'd get some of that, but hopefully in moderate amounts that wouldn't shout down the real discussion.

What will people of the future think of us? Will they say, as Roger Williams said of some of the Massachusetts Indians, that we were wolves with the min
[ Parent ]

The Value of Equity (2.92 / 13) (#6)
by Kwil on Tue Oct 28, 2003 at 03:46:01 PM EST

As you make equity more valuable, there is more incentive to try to game the system. Given that story posts would gain considerable equity, an obvious way to do so would be to create a slew of fake accounts then post a story with your true account and use your fakes to fire it to FP.

While I like the system, I feel that without some kind of account controls, you'd actually end up with a worse K5 than before, as now there's a strange sort of prestige to being able to game the system well enough to then be able to turn around and crapflood it.

That Jesus Christ guy is getting some terrible lag... it took him 3 days to respawn! -NJ CoolBreeze

Probably so (2.50 / 8) (#18)
by localroger on Tue Oct 28, 2003 at 04:11:45 PM EST

without some kind of account controls, you'd actually end up with a worse K5 than before

The whole idea is that you can't really make it impossible to troll or crapflood; setting up insurmountable barriers only invites the really talented 1337 g33k5 to crack their knuckles and go to work on you. The whole idea here is to make undesirable activities expensive rather than impossible.

Basically, if you make the occasional troll or crap-drop something anybody can do once in awhile, the whole thing loses its prestige and it's not worth gaming the system to do it.

What will people of the future think of us? Will they say, as Roger Williams said of some of the Massachusetts Indians, that we were wolves with the min
[ Parent ]

Figure out what you're really trying to say. (2.80 / 10) (#76)
by ramses0 on Tue Oct 28, 2003 at 10:50:44 PM EST

Neat article, Roger... Hopefully this isn't a sign of the "return of the Meta's" story-arc here on K5. :^)

Anyway, in K5, points, equity, rating, etc. they all don't matter.  You're trying to make them matter, but by making them matter, you make them more open to gaming.  Right now you have 1-2 retards who fly around and make noise, but they are outweighed by generic people who "do the right thing", because there's not really an incentive to do otherwise.

What you think you're trying to say is that Dr. Duck is rating too fast.  What you really mean to say is that "Dr. Duck is rating too inaccurately".  Solution #1, is have scoop auto-scale a user's rating to form a bell-curve distribution (well, with only three ratings, maybe a roughly equal distribution).

My rating history: 3, 3, 3, 3, 2, 1
Average Rating: 2.5

Your rating history: 3, 3, 2, 2, 1, 1
Average Rating: 2.0

Dr. Duck rating history: 0, 0, 0, 0, 0, 0
Average Rating: 0.0

Therefore, if I say "3.0", I really mean 2.5 (because my rating "tends to be" 0.5 higher than average of 2.0).  If you say 3.0, you really mean 3.0.  Dr. Duck deviates from the artificial norm by -2.0, which means that all his ratings have an instant +2.0 bonus to them, meaning his 0.0 really means 2.0. (and of course cap mins's and max's at 0.0->3.0, of course).

That's one option.  Option number two is one I like to call "Guess the Rating".  Basically, after a certain point in time, take a snapshot of any one comment's rating data, and figure out how accurate each person who rated the comment was.

Comment Ratings: 3, 3, 3, 2, 1, 0
Average: 2.0

The first 3 were "off by 1" for average ratings.  Not too good, not too bad.  Dr. Duck (with a zero rating) was "off by 2".

Therefore, you reward the people who consistently "guess right" as far as what the end rating of the comment was.  Since Dr. Duck consistenly "misrates" comments, make his ratings worth progressively less and less (this ends up being your version of equity).

People who rate "well" have a scaling factor applied to their ratings, while people who consistently misrate will eventually get their ratings ignored.  

Consider: 1/(1 - $accuracy), with a max and min cap of 2x.

So.  I rate well, I have a "multiplier" of 2.  Dr. Duck rates bad, he has a "multiplier" of zero or .00001 (meaning his ratings don't count).

But in the end, why show mojo, why have numbers attached to comments in the first place?  Hide the numbers, replace them with text ranging from "Good >= 2.2, Neutral >= 1.8, Bad >= 0", and make a note on that person's profile as to what their most recent (mojo) comment rating is.  Actually, making those tags visible makes them gameable (ooh, let me make LocalRoger's rating "Bad", that makes me cool), so we're pretty much S.O.L. until we can solve the "strict identity policing" problem of the internet.

Fun to think about it, though.

[ rate all comments , for great ju
Parent ]

Wonderful! (2.57 / 7) (#82)
by gzt on Tue Oct 28, 2003 at 11:27:36 PM EST

You're promoting a herd mentality. Reward those who rate exactly like everybody else. An unintended consequence is that those who work to unhide unjustly hidden comments will be acting against their interests to do so, while voting to hide hidden comments will have no cost.

Also, what you propose is in no way similar [thankfully!] to Mr. Local's equity.

While I hate meta debates, I think there is no problem with rating too high and quite a problem with rating "too low". My suggestion is that those who give out an average rating below a certain number [say, 1.2] should be ignored.

[ Parent ]

Yes & No. (none / 2) (#108)
by sllort on Wed Oct 29, 2003 at 01:48:44 PM EST

It does invite a herd mentality, though, what is an online community if not a herd? However it does not promote easily hiding comments: people who try to hide comments that the herd likes will be penalized, as will people who try to unhide comments that the herd wants hidden.

In short it automates the triumph of the many over the few. While it remains to be seen whether this is a Good Thing, I think he's done the best job I've ever seen of postulating an effective way of making it happen.
Warning: On Lawn is a documented liar.
[ Parent ]

The point: (none / 1) (#121)
by gzt on Wed Oct 29, 2003 at 04:01:17 PM EST

I don't think induced conformity is what you want in an online community. I may be wrong, in which case, more power to you.

Your explanation says nothing about what I wrote. My point: comments which get zeroed may not be comments which should be zeroed, and many people check the hidden comments to see whether unjustly zeroed comments are zeroed. It happens: relatively innocuous posts are often temporarily rated below 1. The users give them high ratings in order to remove them from the hidden comments pages, but doing so is voting far from the average score of the comment. Under your scheme, they would have less incentive to do so. Indeed, they would have a positive reason not to do so. Do you want this situation?

A community is different from a herd, in my mind, in that in a herd mentality, the particular will, in its expression, and the general will coincide. In a community, the idea of "general will" is balderdash. Stuff it, Mr. Rousseau.

[ Parent ]

No, I don't want it. (none / 2) (#124)
by sllort on Wed Oct 29, 2003 at 05:29:31 PM EST

But if the object is to create a herd, which many including the author of the story seem to be in favor of, I argue that this is an excellent way of actually implementing it. And yes, eliminating "ratings steering", which is a way in which one voter raises his or her voice to a greater volume than the others by exagerating the rating - the proposed system would have exactly that effect.

And again, no, I don't want it. But I am compelled to point out that it is, unlike 99% of the half baked crap in this article, "actually a functional system".
Warning: On Lawn is a documented liar.
[ Parent ]

Yeah, it is (none / 1) (#128)
by dennis on Wed Oct 29, 2003 at 09:18:29 PM EST

It's already a majority-rules system. There's a single aggregate rating, which theoretically reflects the collective judgement of the site's users. The above system simply does this more effectively. I think it's the best idea I've seen here thus far.

You can greatly reduce the "acting against my interests" problem if the multiplier has no effect except on your rating effectiveness, and not any other privileges. There's no point in having powerful ratings if you're not willing to use them to disagree with others' ratings.

If you want a site that is not a herd, a simple adjustment to this system will do it for you. Instead of multiplying raters by how well they correlate with the average, multiply them by how well they correlate with you. Same sort of stats used by collab filtering sites everywhere. Now, you have a strong incentive to rate like you really think, because it will affect your view of everybody else's ratings. And you can still have the global value, to give a meaningful number to the lurkers...and the global value is even more accurate.

All of which is fairly meaningless if you can't hide all comments below a threshold you set yourself, imho.

[ Parent ]

Nice (none / 2) (#109)
by sllort on Wed Oct 29, 2003 at 01:53:21 PM EST

That's the closest thing I've ever seen to a ratings system that might actually accomplish its intention: allow the vast majority to triumph over an aberring minority. It would be very hard to defeat, especially if new accounts were given an "ineffective" rating bonus to start out with which they could only improve through agreement.

I'm not sure the goal is worthy, but your implementation is golden.

The bad news is that once a comment has acquired about 10 "good" ratings, many will be scared to disagree. Imagine a comment which links to an image relevant to the article on the poster's server. After ten good ratings, the image is switched with something heinous. Who at that point has the guts to zero the comment? Of course, just disagreeing on one comment should only affect your, um, Kar-jo a little, I suppose.
Warning: On Lawn is a documented liar.
[ Parent ]

-1 Article may have a few worthwhile points (1.83 / 18) (#8)
by Tex Bigballs on Tue Oct 28, 2003 at 03:48:45 PM EST

but it's getting to the point where K5 should just change the tagline to "technology and culture meta from the trenches" because lately it seems the only thing that's discussed around here.

To be fair, Tex (2.12 / 8) (#10)
by A Proud American on Tue Oct 28, 2003 at 03:53:09 PM EST

... the last four articles that I can see as having been posted are:
  1. Something from you about gaming
  2. Something about CSS
  3. Something from me about homosexuality in schools, and
  4. Another article by me about Russian alcohol problems
Now, unless Rusty turns gay (see 2) and turns K5 into a gaming site (see 4) using nothing but CSS code (see 3) after a drunken stupor (see 1), I hardly see how all the recent articles are Meta.

But alas, IHBT (-;

The weak are killed and eaten...

[ Parent ]

You're conveniently forgetting (2.25 / 8) (#16)
by Tex Bigballs on Tue Oct 28, 2003 at 04:04:05 PM EST

ubernostrum's meta article that was in the queue for about 2 whole days before it ended up getting dumped during the same time as #4. That was less than a week ago and already we're getting more meta dumped into the queue.

[ Parent ]
Discussed, or posted? (2.40 / 5) (#12)
by Rich0 on Tue Oct 28, 2003 at 03:56:32 PM EST

There is plenty discussed.  However, it seems like anything liberal gets booted by the conservatives, and anything conservative gets booted by the liberals.

If you post on vertex shaders you're probably safe.  However, anything related to the world at large or politics is cannon fodder.

I don't see why we can't discuss the pros and cons of a controversial topic rather than feeling the need to down-mod or dump anything that disagrees with one's personal opinion.  If I 0'd anything that I disagreed with, I'd be giving out a lot of 0's.  In fact, I am about as likely to 0-1 something I disagree with as I am likely to 0-1 something I do agree with.  If the comment in meaningful and informative, then I think it is worth my reading, and the reading of others.  Just because I hold a certain political view doesn't mean I think it's smart to put on blinders...

Unfortunately, I think a lot of users here would disagree.  For some, mods are a great way of saying "no!  yes!  no!  yes!  no!" and whoever has the majority wins.

[ Parent ]

It's not rally discussed that much (1.57 / 7) (#13)
by iasius on Tue Oct 28, 2003 at 03:57:21 PM EST

most of the posts in meta articles just say "-1 meta" or "not another meta article".
Once they're out of the editing queue they are usually quickly voted down.
Either k5 readers mostly don't want any change, or a minority quickly uses every means available, so change doesn't happen.

the internet troll is the pinnacle of human evolution - circletimessquare
[ Parent ]
True (1.66 / 6) (#17)
by Tex Bigballs on Tue Oct 28, 2003 at 04:05:25 PM EST

you would think people would get the point

[ Parent ]
Nah (1.40 / 5) (#31)
by sllort on Tue Oct 28, 2003 at 06:40:55 PM EST

They're just waiting for an entry from the local popularity contest winner.

Man if Rusty wanted to hide information from users, he shouldn't hide rating totals, he should hide the identity of the author until the story's posted, and delete any comments that leak the identity.

Sure, it's a doomed enforcement scheme at its worst. But at least it would keep the story queue from reminding me of voting for prom queen.

This story is flawed in its factual premise, reasoning, and conclusions, but it's gonna get voted up like a rocket all the same.
Warning: On Lawn is a documented liar.
[ Parent ]

Yeah... if you or me wrote this shit (1.71 / 7) (#33)
by Tex Bigballs on Tue Oct 28, 2003 at 06:42:34 PM EST

it would be voted down to the center of the earth

Oh well, the way I look at it is, at least his house caved in and his teeth are all jacked up

[ Parent ]

If you're not part of the solution... (1.66 / 6) (#29)
by skyknight on Tue Oct 28, 2003 at 06:33:41 PM EST

then you're part of the problem! I specifically remember you +1FPing my latest meta article, perhaps for selfish reasons. ;-)

It's not much fun at the top. I envy the common people, their hearty meals and Bruce Springsteen and voting. --SIGNOR SPAGHETTI
[ Parent ]
Hahaha yeah this is true (2.00 / 9) (#30)
by Tex Bigballs on Tue Oct 28, 2003 at 06:39:08 PM EST

fair point indeed.

But your article was a pretty uncontroversial criticism over one of the eccentricities of the code. This is yet another wanky piece about how to solve trolling/crapflooding/etc.

[ Parent ]

meta meta....brilliant <nt> (1.50 / 6) (#90)
by auraslip on Wed Oct 29, 2003 at 03:23:04 AM EST

[ Parent ]
meta meta meta....brilliant (none / 1) (#130)
by scanman on Wed Oct 29, 2003 at 10:06:22 PM EST

If only I could turn "meta" into "meat" I would be very fat indeed.

"[You are] a narrow-minded moron [and] a complete loser." - David Quartz
"scanman: The moron." - ucblockhead
"I prefer the term 'lifeskills impaired'" - Inoshiro

[ Parent ]

Why didn't you mention Perl Monks/Everything? (2.61 / 13) (#11)
by Night In White Satin4 on Tue Oct 28, 2003 at 03:55:01 PM EST

Either way, I think that the best advantadges of K5 over Slashdot are that it's unregulated. I.e., since it's edited by the people themselves it tends to cater to that need much better, as opposed to a closely edited site such as /.. I don't like you suggestions of creating an economy and having to spend "equity" to colaborate since that would, effectively, discourage colaboration. By collaboration I mean moderating also. Who cares about drduck? Let him, if that's his kink. The modbombers get their ratings revoked as it is.

I think PerlMonks/Everything's system is pretty good at achieving the stated goals: you put a post, it gets rated up which in turn give you XP. However, I think it has two very big flaws: some people with monstous XP which, though having little special priviledge, stand out. I don't think any users should stand out for any particular reason.

The other flaw is that it focus the discussion too much. That's great for Perl Monks because their discussion is centered around one topic — Perl. That would be awful to K5 because it'd lead to wars between groups with different interests, and one squashing the other.

Perhaps we need to ask ourselves what we would want to achieve with the proposed changes and all I can think of is censoring. While it might weed out some of the idiots, even said idiots, after a little while, start to behaving themselves. Even rmg is a cute little sheep now. I think that any big changes that might alter the liberal feeling of the site would do more harm than good.

All in all, I think K5 is the best it can be and I'd introduce no big changes. The comment system is pretty good (I liked the changes) and maybe I'd change a detail or two.


Two things about everything2 (2.80 / 5) (#23)
by mcc on Tue Oct 28, 2003 at 04:34:47 PM EST

E2's system is a tiny bit more complicated in a couple of interesting ways.

One, the thing NIWSX does not mention is that the more XP you have, the more ability to rate you have. Sort of. You gain "levels" within e2's system based on how many posts you've made and how much XP you have. You must be level 2 to rate posts, and with each level you get more "votes", each vote letting you rate on one post. Once you reach level 4, you recieve an ability called "Chinging", which will allow you to mark a post as exceptional. When you do this, it is displayed on the front page on the "Cool nodes" list for a short while, ensuring it gets lots of exposure and gets rated way up.

Two, the thing is, on e2, XP really is not very useful. Almost nobody has XP as their limiting factor once they get past level 3 or so-- generally you have XP to spare and the thing you need to level up is to write more nodes. So to this extent XP mostly serves as a check-- basically XP only comes into play if you've been writing actively bad nodes. If you have, people rate you down, and your XP limits you from leveling. So XP is more of a block on bad behavior than a reward for good behavior.

This is all complicated a bit because there's something called "Merit" that was recently instituted that bumps your level up if you have an abnormally large number of high-rated nodes, with the hope being this will reward people who write a small number of good nodes rather than a large number of mediocre ones.

Aside from that, the absurd meta-wankery of k5er-quoting sigs probably takes the cake. Especially when the quote itself is about k5. -- tsubame
[ Parent ]

localroger has good points tho (1.75 / 4) (#97)
by phred on Wed Oct 29, 2003 at 10:38:31 AM EST

K5's moderation isn't very effective. However, K5's software change methods pretty well involve no testing and typically screw the site up for weeks, thats why I think K5's moderation should not change anymore, because the "CMF" has zero programming resources.

[ Parent ]
Flawed assumptions & reasoning, but fun stuff (2.87 / 16) (#26)
by sllort on Tue Oct 28, 2003 at 06:01:15 PM EST

But I still love all Meta stories! So here goes:

Whether you call it Mojo, Karma, "Standing," or something else, all content rating feedback systems have some sort of currency.

There should really be some exploration of why this is. This really only needs to occur because we don't have a good way of stopping people from having multiple identities. Sites that only accept members with credit cards don't need this at all because they have authentication. Each person gets one account, and it's enforcable, at which point this entire economy system never has to come into being. People are very polite on sites like this. Over at fark, only paying members can become a member of the hand-picked goon squad of moderators, who at that point have the ability to enforce permanent or two-week IP bans. After the first 3 or 4 "Internet executions", fark's noise problem essentially stopped.

On the other hand the Slashcode system is pretty hard to game

Incorrect. The Slashcode system is unbelievably easy to game, but the people gaming it want to ensure that the people writing Slashcode don't catch on and fix it. Hence, it gets hard to detect to the average observer as well. Note that unlike on K5, where user-created discussions are frequently moderated, mod points are never used in user-created discussions on Slashdot due to scarcity of points. So comparing K5 and /. in this way is pretty flawed, since /. only works for rating comments in a story queue driven by a commercial editing staff. K5 is a system for moderating user-created content, while Slashcode is a system which only works for moderating administrator-driven content. Hence, Scoop has a much bigger task ahead of it. Slashcode's fabled "anti-crapflood" technology actually doesn't work at all in user-created discussions, which to this day still see scripted spam attacks in the 3-4 thousand comment range.

Equity should be worth something, so that one has an incentive to use a single account and keep it in good standing.

This will cause what we have already seen, the Jekyll/Hide syndrome. Every "troll" has another account, their do-gooder account, from which to access all the information hidden from the troll account. Hell I did this just to make sure that nothing would ever be hidden from me by modbombing scripts (by the way, who used the word "modbombing" before me? i suspect i will never get credit).

Poor equity isn't just a number; you notice it in terms of lost conveniences before you're effectively kicked off the site.

Don't be silly. You meant to say: "Poor equity isn't just a number; an account may notice it in terms of lost conveniences before the account is effectively kicked off the site, forcing the owner to generate ten more."

So, here I will list all your ideas for how a user should be able to gain equity:

  • Leaving highly rated posts should grow one's equity, and leaving poorly rated posts should shrink it.
  • It should go without saying that getting a story voted up should be worth a healthy shot of equity, and getting it to FP should be worth even more.
  • Conversely, the site might sell equity outright for dollars under reasonable terms discouraging abuse.
  • In order to "charge up" the economy it might be worthwhile to dispense equity simply for loading pages (at a low, humanly reasonable, capped rate) to encourage the "silent majority" of lurkers to participate

Equity from Ratings: What's to keep people (let's call them "trolls") from creating their own discussions and moderating themselves up, thereby becoming the Richest In The Land? What's to keep them from creating secret self-reinforcement discussions? I ask you this truly.

Equity from Stories: The true end of any attempt to make story selection objective, this would be the last step into the inferno. I believe the case has been more succinctly and elegantly stated here.

Equity from page loads: Do you seriously want to make drduck a rich, rich man? If ever there was a formula to bankrupt Voxel.net and bring about the triumph of the, um, "trolls", this is it. By all means, go for it.

Equity for cash : Well, as a CMF board member, I see no objections to this one. Tally ho.

So, in short, you have accurately described the problem (no working equity system) and failed to come up with a viable solution. But you have definitely thought about it a great deal, and the discussion is worthwhile. For what its worth, I can't think of a single way of dispensing equity that the most determined users could not game to the disadvantage of the rest. What you really need is a "content detector". It's a tough problem.

There should be sanity checks. There is no reason at all to allow an account like drduck to moderate faster than anyone can read.

This is, properly implemented, a good idea. Hard to do right, but it couldn't hurt. I kind of like the anti-spammer tactics where a link which no human could see (or follow) would lead to an infinite tarpit of garbage stories and bogus comments, which the script would traverse and rate, thereby revealing itself and getting its IP banned. But I digress.

My solution? Negative voting attracts negative voters. You shouldn't be able to vote against comments or stories or anything else. I wrote some of these ideas down here. I don't claim to be any more correct than you, but there are some ideas in there to kick around as well.

Still voting up your story as +1: promotes discussion cause I'm a sucker for the Meta section.
Warning: On Lawn is a documented liar.

A few minor points: (1.28 / 7) (#78)
by rmg on Tue Oct 28, 2003 at 10:59:16 PM EST

Slashcode's fabled "anti-crapflood" technology actually doesn't work at all in user-created discussions, which to this day still see scripted spam attacks in the 3-4 thousand comment range.

It is true that scoop manages to prevent attacks of this sort, but only by crashing. Posting large numbers of page lengthening comments is just one of many ways to make K5 go down like an altar boy.

Every "troll" has another account, their do-gooder account, from which to access all the information hidden from the troll account.

This is not true. I have no accounts that I use for anything other than trolling.

Equity from page loads: Do you seriously want to make drduck a rich, rich man? If ever there was a formula to bankrupt Voxel.net and bring about the triumph of the, um, "trolls", this is it. By all means, go for it.

It is best not to point out the absurdity of ideas like this one. If the local genious, Local Roger, is not able to think this through properly, it's a safe bet rusty won't be any more capable.

_____ intellectual tiddlywinks
[ Parent ]

OOh OOh. Neat new feature idea. (none / 2) (#133)
by gte910h on Thu Oct 30, 2003 at 10:28:51 AM EST

Over at fark, only paying members can become a member of the hand-picked goon squad of moderators, who at that point have the ability to enforce permanent or two-week IP bans. After the first 3 or 4 "Internet executions", fark's noise problem essentially stopped.

Rusty should get a credit account with someone. Accounts by default start off as unvalidated, and cannot moderate, but if they allow rusty to verify their existance by checking their address on their credit card, then they can moderate.

Anyone who makes a donation through an account should also get this privledge. Basically any time rusty gets someones address, they should be able to moderate. Rusty shouldn't keep the users CC#, but he should keep their addresses.

I believe this is what many pornography sites purport to do, but they instead charge your card for mucho dinero. I trust rusty not to do that.

[ Parent ]

That would change everything (none / 2) (#137)
by sllort on Thu Oct 30, 2003 at 12:57:18 PM EST

It would change the free & open character of the site.
It would vastly reduce the pool of moderators.
It would allow rusty to actually effectively ban people from moderating, which could be good or bad.
It would also, I am sad to say, work quite well.

But at what cost?
Warning: On Lawn is a documented liar.
[ Parent ]

Lessons from the market? (none / 3) (#143)
by sane23 on Fri Oct 31, 2003 at 01:16:05 AM EST

Heck, why don't we try slavishly applying the techniques used by financial markets to see if we can gain some insight.

Gaining equity from posts:

Let's view posting a comment as making an investment. You pay equity corresponding to the middle of the scale to post, and as your rating rises or falls, your unrealized equity also rises and falls. When equity is needed for some other use, the system can arrange a "sale", converting unrealized equity to realized. For simplicity, this might instead happen after some time or lessened activity threshhold.

This rewards posters of comments that rise, remains neutral for comments that are uneventful, and helps to prevent future posting by accounts whose comments are downrated. This only suffers from the "self-congratulatory community" problem if equity is easily available to crapflooders. I propose a solution to that problem at the end of this comment.

Incentives for accurate rating:

Having rating cost equity is a good idea, but perhaps an incentive can be given to those who rate comments in such a way that reflects the general consensus. Rating a comment can affect the value in the standard averaging manner, but raters can gain equity by uprating a comment whose value is "undeservedly" low and waiting for the rating to rise as others do the same. The converse is also valid, as raters downrate comments whose value is "undeservedly" high. Those who "bet" incorrectly can certainly lose equity.

Gains and losses from ratings should probably be proportionately worth less than those from posting, to ensure that the site doesn't turn into a nest of futures traders, but a minor incentive could still be worthwhile in producing quality ratings.

Gaining equity initially:

If new accounts are granted a certain amount of newly minted equity, then abuse is nearly guaranteed. How do people initially acquire equity in the real world? They get a loan. Unfortunately, we can't track a "credit rating" because of the new account problem. So, we have to use the old fallback favored by conservatives everywhere: inherit it.

If an initial equity stake could only be acquired by physically contacting a current user and arranging a transfer, or possiby a loan (interest-bearing or not), then the community would hopefully become a bit more connected and certainly more resistant to newcoming crapflooders. This is similar to LiveJournal's system of referrals, but in that system, one bad referral has unlimited access to crapflood forever. With this system, they are limited by their initial stake, and the lenders can apply whatever screens they may choose. If they make a bad referral, they can accept the loss and respond as they see fit.

A downside: this barrier to entry could prevent new users from joining, or could make the community more insular. Time will tell on that one, I suppose.

Some of this stuff could apply to stories and diaries, but I haven't thought it through yet.

[ Parent ]

+1 FP (2.69 / 13) (#27)
by Prophet themusicgod1 on Tue Oct 28, 2003 at 06:03:09 PM EST

<u>first the bad</u>
  • gaming users as noise. There are always going to be nutcases, morons, perverts, republicans, trolls, "liberal socialist faggot hippies" and people-who-put-terms-on-people-too-readily and people-who-put-lines-in-between-complex-words-they-aren't-smart-enough-to-look-u p-the-right-word-for... this is part of moderation. rusty pointed out before i got here that given enough moderators, the odd rating is eventually modded out into oblivion. we need to encourage moderation, which is what you are doing here, but there is no outright killing this behavior off. this is why when unlimited moderation happens, it was a good thing, compared to non-unlimited moderation.
  • 5. [-1] [0] 1 2 3 4 5 [6]. it makes _sense_ you have a scale. i could see 2.5, pi, and sqrt(2) working, too. why? because it is imaginable on a scale. hell, even a -1,0,1 where you give a decimal value would be ok... why? because you can picture the best comment, you can picture the worst comment, and generally comments are going to be between these two extreems(although they actually tend to not---explained later in post). the 1-5 rating system had something at least i could wrap my mind around, and it made coming here that much more interesting.
  • ignore ratings. you see, moderation _does_ have an effect in K5. i set my sort-based-on-ratings switchthing to "ignore ratings"...and boy can i ever tell the difference. sure, the brilliant, inspired thought is still there, but it is weeded in between trolls, porn, and spam. i don't mind trolls and porn, but i prefer them to be near the bottom of my page. yes, i know, i could still sort by ratings, but this doesn't make sense to me, being rated all under 3. what is a medium comment? 0 ?
  • new user equity leads to throwaway user accounts. this is basically what we have now, new accounts are valuable first without contributing. your equity solves this, saying you must earn, through contribution, the ability to moderate.
  • extreem ratings BECAUSE of extreem comments it has been said again, and again, and again, that people here tend to vote mostly in 1's and 5's, with little in between. rather than chalk this up, like most people here, to laziness and eagerness to vote things up or down, i chalk it up to the comments. there have been SOO MANY ingenius posts that have just blown my mind. i remember more than a few times rating every other comment in a post down one or two notches, because they just missed the point that was so imperative to the story, that only one person seemed to get. and then realizing someone else had an even greater, beyond beleif, idea than he had. on the flip side...there is a lot of spam, and junk. you know the type... 'FOR PENIS ENLGUGEMNT SUCK MY COCK LINUS TROVALDS'. some people just have too much time on their hands. your system solves this, to an extent though. while we would not be taking their right to post such drivel away, we would discourage it, and make it more difficult.
  • negative moderation. i think either a) you should not be able to destroy equity(via negative ratings) or b) you should only be able to destroy equity at a very, high cost. why? because this is just another step towards our system commiting suicide. it shouldn't ever shoot itself with it's own energy. if you are pissed off at a person, there needs to be some other way of venting, otherwise wars will be easier to start, the system will crumble faster because it's own momentum will tear it apart. most slashdot moderations are negative, correct?

<u>now for some ideas</u>
  • post more than one (x) per minute, for the cost of X equity YES this means more trolling, but if you set the equity costs high enough, you could easily recoup anything lost with that.
  • equity cost stable, the amount of equity given out slowly increased or decreased, depending on financial state this made alot more sense 5 minutes ago.
<u>the good</u>
  • equity for money you sucessfully took down two main problems with this site in one stroke, well done
  • low latency for equity- brilliant
  • moderation for equity
  • diaries for equity -great idea
  • bestowing_equity for equity this may be the most important part. because there becomes an ecconomy created, and most of the kinks will iron themselves out. there is reason for people to contribute, ala so they can contribute more. it's a great feedback system, really, and can only promote good things, or is heavily biased towards these good things.
  • timing because the current system sucks

<u>and now for some initial replies</u>
  • Tex: First, we need a god-damn working forum. rusty screwed the one we had up, and this is the turmoil/dialectical counter-reaction.
    Then we need to repeal the anti-terrorist/martial law-laws in canada, the united states, and really anything else of the sort elsewhere in the world. or hell, even in one country so we can all move there so we aren't sent to the death for doing something like 'reading kuro5hin'.
    then and only then can we really go back to technology and culture. although, perhaps we don't need this for the 'trenches' bit, unless that was the point and i'm missing it
  • pmc: drduck _does_ contribute. his contribution's are not interpreted correctly, and tend to be felt as noise, but they are contributions nonetheless. his viewpoint has forced us all to think about ourselves, and revaluate our opinion of him, and the k5 rating system as a whole. this system actually not only allows drduck to continue, and embraces such things, but also solves the problem he may pose to you. which is why this system is so great.

address these comments and i may consider voting 'equity' instead of 'mojo'.
"I suspect the best way to deal with procrastination is to put off the procrastination itself until later. I've been meaning to try this, but haven't gotten around to it yet."swr
One comment (none / 3) (#127)
by Valdrax on Wed Oct 29, 2003 at 06:33:56 PM EST

As someone who regularly reads and meta-moderates Slashdot, I'd have to say that the VAST majority of moderations there are positive.  You may see more hidden comments than positively moderated ones in some articles, but the vast majority of those hidden comments got ~1 negative mod while the highly moderated comments tend to get ~2-3 positive mods.  However, I'd say that there are less hidden comments than positively moderated ones in most articles outside the BSD section.

[ Parent ]
circletimessquare's tired wheeze (by now) (1.88 / 9) (#32)
by circletimessquare on Tue Oct 28, 2003 at 06:42:17 PM EST


removes all of the asocial negative vibes form the rating system

removes ideologically driven asocial voting

makes everybody shiny happy smiley

cleans and brightens teeth

fully endorsed by turmeric (seriously, look at the comments on my diary link below)



The tigers of wrath are wiser than the horses of instruction.

Well just think circle (1.57 / 7) (#34)
by Tex Bigballs on Tue Oct 28, 2003 at 06:53:54 PM EST

if you wrote a couple hundred words on this subject, used some catchy buzzwords like "equity" and "economy", put it all into neatly nested bulletpoints, and kinda just bullshitted some half-baked ideas for filler you could very well have a front page meta article on your hands.

[ Parent ]
You forgot step 5: (1.66 / 6) (#36)
by sllort on Tue Oct 28, 2003 at 06:59:06 PM EST

5) Hacked a popular account to post it from.
Warning: On Lawn is a documented liar.
[ Parent ]
There are ways (2.33 / 6) (#40)
by localroger on Tue Oct 28, 2003 at 07:07:22 PM EST

...to get a popular account, or so I've been told. Other than guessing the password of someone who's done it before, you might try posting stuff that people actually want to read, a little bird tells me that that's a workable scheme.

What will people of the future think of us? Will they say, as Roger Williams said of some of the Massachusetts Indians, that we were wolves with the min
[ Parent ]
Aw, you think I'm bustin on you. (1.50 / 4) (#43)
by sllort on Tue Oct 28, 2003 at 07:16:06 PM EST

I'm not. I like your fiction.
And before you go lecturing me about how to abuse multiple accounts, ask yourself, how do you know I don't?

Warning: On Lawn is a documented liar.
[ Parent ]
Mmm Hmm. (2.28 / 7) (#87)
by RobotSlave on Wed Oct 29, 2003 at 12:19:13 AM EST

"posting stuff that people actually want to read"

What you are suggesting, of course, is that the surest route to popularity is pandering to the biases of your audience. I don't doubt that, but I would humbly suggest that subscribing to those biases oneself, and thus being perhaps unaware of the pandering, would make the whole enterprise quite a bit more palatable.

For those of us who are cursed not only with differing views, but with integrity as well, however, life isn't quite the facile after-school special you make it out to be.

[ Parent ]

Not so bad. (none / 1) (#129)
by dennis on Wed Oct 29, 2003 at 09:29:50 PM EST

Personally, I rate comments up that I disagree with, as long as they're argued well. If most people disagree with what you say, work extra hard on your arguments. Who knows, maybe you'll even convince somebody.

[ Parent ]
No, not necessarily. (none / 1) (#140)
by RobotSlave on Thu Oct 30, 2003 at 04:12:46 PM EST

I don't have much trouble getting articles posted, when I actually make an effort. But I do have to make an effort. I don't usually suffer fools gladly, so it takes a bit of effort, and a temporary abandonment of principle, for me to pander to them.

People like localroger don't get very far when they tackle the actual arguments of people who disagree with them, so they instead resort to attacking the person, or the delivery. I don't think much of such efforts, but then I'm not a self-described genius like localroger, so maybe I'm missing something subtle.

[ Parent ]

really? (none / 1) (#153)
by Hakamadare on Thu Nov 06, 2003 at 12:32:47 PM EST

I don't have much trouble getting articles posted, when I actually make an effort.

hm, i'm not so sure that's the case.  it appears that you don't care to "make an effort" very frequently - you don't have enough data points to be able to make that assertion.

your pedal error story was entertaining and useful, but other than that, you seem to be an arrogant windbag, prone to ad hominem attacks and bitchiness.  for you to pass that off as having anything to do with "principle" is just priceless, especially when you complain about others doing the exact same thing.

Schopenhauer is not featuring heavily on the "Review Hidden Comments" page at the moment. - Herring
[ Parent ]

haha (none / 0) (#155)
by ColeH on Sun Dec 07, 2003 at 02:18:20 AM EST

You are a riot. Did you ever decide what cheating is?? Dude, you need help - seriously.

[ Parent ]
ah yes, the familiar straw man (none / 2) (#135)
by Hakamadare on Thu Oct 30, 2003 at 10:41:29 AM EST

those of us who are cursed not only with differing views, but with integrity as well

it's so gratifying to portray yourself as the principled idealist, struggling gamely upstream against the envious scorn of the sheeplike masses who Just Don't Get your genius, isn't it?  a veritable Randian masturbatory fantasy, indeed.

and it's so much easier to accept than the possibility that, just perhaps, a good route to becoming a "popular user" (whatever the hell that means) might have something to do with not being a complete and utter asshat.  believe it or not, it is possible to disagree with the status quo without being obnoxious; what's more, just because you're obnoxious doesn't automatically mean that you disagree with the status quo.


p.s. the truly sad thing is how little time it took to find the links in the above paragraph.  perhaps you've just been in a pissy mood recently, eh?
Schopenhauer is not featuring heavily on the "Review Hidden Comments" page at the moment. - Herring
[ Parent ]

See what I mean? (none / 1) (#139)
by RobotSlave on Thu Oct 30, 2003 at 03:51:56 PM EST

[ Parent ]
Again, I agree. (2.00 / 3) (#35)
by sllort on Tue Oct 28, 2003 at 06:58:15 PM EST

Again, I pile on:

Warning: On Lawn is a documented liar.
[ Parent ]

to offset the rich bastard problem, (1.83 / 6) (#42)
by Kasreyn on Tue Oct 28, 2003 at 07:12:00 PM EST

gain in equity from a comment being well-rated could be made lower than loss in equity for posting a comment or making a rating (in terms of crapflooding). Thus two crapflooders working together could only lose equity, not gain it. However, a crapflooder with say 5 or 6 accomplices with equity to burn, could cause his crapflood to be well-rated, however, the overall loss in equity amongst them would be huge.

The general idea: Posting a comment costs 5 equity. Rating a comment +1 costs 3 equity. Having a comment of yours rated +1, gives you one equity. (thus, 66% of the equity spent on rating "evaporates")

Since this would result in the total equity in the community bottoming out, I'd suggest a monthly infusion of equity, equally distributed between accounts.


"Extenuating circumstance to be mentioned on Judgement Day:
We never asked to be born in the first place."

R.I.P. Kurt. You will be missed.
Wow. (2.33 / 6) (#44)
by sllort on Tue Oct 28, 2003 at 07:17:31 PM EST

If this plan gets anywhere near implementation, please drop me an email reminding me to write a script to generate 1,000 accounts from which to mod myself up.
Warning: On Lawn is a documented liar.
[ Parent ]
Stanford Rating System (2.92 / 14) (#47)
by dennis on Tue Oct 28, 2003 at 07:30:42 PM EST

Your proposal after "Extreme rating might cost more" reminded me of the Stanford Scale for rating the attractiveness of women, which I read about Omni magazine a long, long time ago.

The guys at Stanford considered the usual one-to-ten scale to be both subjective, and unstandardized. If you give an eight rating where I give a six, it could be that we rate on different scales, or that we had different subjective responses to the individual in question. Comment rating poses similar difficulties.

A more rational scale, they reasoned, would give an objective rating of the subjective response itself. The Stanford Scale ranges from 2 to -2, with clear definitions as follows:

2: You can't stop looking.
1: You can stop looking, but you'd rather not.
0: It doesn't matter whether you look or not.
-1: You can look, but you'd rather not.
-2: You can't look.

Just how to modify these definitions to apply to comments escapes me thus far, but this scale does happen to exactly correspond to your proposal to charge points for ratings.

We've already got it (3.00 / 6) (#110)
by sllort on Wed Oct 29, 2003 at 02:08:08 PM EST

The label on zero says "Hide". It means, "this comment is so abusive it should be hidden from view", i.e. "-2: You can't look.".

The problem is, 0 still gets abused, it still gets applied to comments that are neither spam nor abuse, it gets smattered all over, and its most frequent meaning is "0: I am angry".

No matter how much instruction you give people on what the ratings scale means, in the end, the worst rating will always mean "I am angry". Consider Slashdot:

-1 Troll (I am angry)
-1 Overrated (I am angry and cowardly)
-1 Flamebait (I am angry)
-1 Redundant (I am angry and sarcastic)

The only real solution is to remove any outlet for anger from the comment system, and make the most negative reaction to a piece of content be that you ignore that content. Ignore bad stories, ignore bad comments, ignore things that make you angry.

Cut the loop!
Warning: On Lawn is a documented liar.
[ Parent ]

what do you want to promote? (2.50 / 4) (#49)
by martingale on Tue Oct 28, 2003 at 07:59:13 PM EST

What you're proposing is to give users extra responsibility to manage their allocated currency, in the hope that managing both their currency and their comment production together is a task which generally obtains "good" behaviour.

Stated in this way, I'm not so sure I see that you've given compelling arguments on why "good" behaviour is encouraged by the extra complexity.

There are two ways of arguing that an improvement is effected. One is to show that managing your currency actually improves the quality of what you produce (where you produce either comments or comment ratings). The other way is to argue that managing your currency effectively reduces production failures (ie you produce fewer "defective" comments or comment ratings). I think you've mostly argued the latter here.

Stated in this way (if you accept my description above), I think I can restate your argument as follows: if the "bad" people spend a lot of time and effort managing their currency, then they aren't spending this time and effort producing bad ratings and comments. In other words, give them a new toy to play with, and they'll leave ours alone. Of course this assumes that 1) they care about currency and 2) they have limited time and effort to spend.

I don't think I'm convinced. Both 1) and 2) are generally false, unless I'm missing something.

Reason for the complexity (2.25 / 4) (#50)
by localroger on Tue Oct 28, 2003 at 08:04:58 PM EST

The purpose of the complexity isn't to "confuse" those gaming the system. It is to provide robustness. Generally, you can only game one parameter at a time, just as only one feedback mechanism tends to want to collapse at a time in an ecosystem or economy. The idea is to spread the equity distribution mechanism around so that a hit here or there doesn't disrupt the basic purpose of the system.

Basically, the idea isn't really about any of the specific ideas I mention; some of those may have merit, some may not, as various comments have pointed out. The main point is to have a lot of different feedback mechanisms operating at the same time so that gaming one does not give the gamer keys to local reality.

What will people of the future think of us? Will they say, as Roger Williams said of some of the Massachusetts Indians, that we were wolves with the min
[ Parent ]

complex systems (2.50 / 3) (#57)
by martingale on Tue Oct 28, 2003 at 09:05:29 PM EST

The purpose of the complexity isn't to "confuse" those gaming the system. It is to provide robustness.
Fair enough, but I'm not sure that complexity implies robustness. Often, complex systems can be unstable too. It is true that complex systems which have survived for a while are often robust, simply because if they weren't robust, they wouldn't have survived.

As an example of a robust system, currency in the real world is arguably robust. I say arguably, since all examples I can think of also employ wealth redistribution methods to prevent collapse or chaos.

So assuming robustness, duplicating currency on k5 should imbue to the k5 system the robustness of the real world currency system.

The main point is to have a lot of different feedback mechanisms operating at the same time so that gaming one does not give the gamer keys to local reality.
I've now convinced myself that doing so is a bad idea. To summarize my thoughts: by introducting complexity, there is no guarantee that the system will become more robust or less. One would have to copy an existing robust system to inherit the robustness. On the other hand, by introducing complexity, we deliberately make life more difficult for casual k5 users. This may or may not be a good idea, although I tend to think it isn't a good idea.

However, it may well be an interesting mental exercise to map k5 on to some simple, well understood system such as a predator-prey model. Food for thought.

[ Parent ]

Currency exchange (2.50 / 6) (#52)
by dennis on Tue Oct 28, 2003 at 08:07:53 PM EST

So, you spend one or two points to rate a comment good or bad. And if one of your comments gets rated good, you gain points.

It makes sense, therefore, that when someone spends two points to rate your comment good, those two points should go to you.

But if someone rates your comment one or two points bad, you should lose points. But they lose points too. Where do the points go? You don't want to deflate the economy.

One answer...the points go to the next poor schmuck who reads the comment. He deserves it as compensation for his pain. And if you have the ability to hide low-rated comments, this gives people an incentive to spend some time in the low-rated comments, so you avoid the slashdot problem of good comments languishing in the doldrums. If a rater disagrees with the low rating, he can give the same points he just earned back to the comment in question.

The implementation prevents a bit of a problem, since normally we bring up a lot of comments on one screen, instead of viewing each individually (which is much less convenient). But this isn't hard to deal with; just hide all comments that have a negative balance, and for those comments alone, give people an option to view them one at a time, at random, in exchange for one point apiece. Note that the comments maintain their low ratings unless the viewers give the points back to the comments; the points given to the viewers are sitting out on the porch, and only the viewer can choose to put the point back in the comment's safe.

This doesn't quite work, because if the points on the porch run out, you don't have any more to give to the viewer, and presumably would no longer show the comment to anybody...so a below-threshold comment that inspires every viewer to rate downwards would actually get more pageviews than one which dips a bit negative and is neutral from then on. These especially horrible comments would therefore be the most fit, and over time the surviving low-threshold comments would become a cesspool of the worst comments ever written on the site.

But this is fixable as well. It should work if you a) don't allow negative ratings on below-threshold comments, and b) Don't hide comments until they reach, say, -10, and don't make them visible again until they get all the way back up to 0. That way comments aren't constantly popping in and out. A comment isn't hidden until 10 people rate it down, and it stays in that mode until ten more people view it without giving any points back...but if the majority of people viewing it do give points back, it will claw its way back into general visibility.

All this isn't quite consistent with the idea of different users setting different thresholds of visibility...but this comment is long enough.

Checkmate! (none / 2) (#119)
by sllort on Wed Oct 29, 2003 at 03:30:16 PM EST

just hide all comments that have a negative balance, and for those comments alone, give people an option to view them one at a time, at random, in exchange for one point apiece.

Ok, say I'm a "troll", my pet drduck script reads these comments until my equity tank is full every night at midnight. Now I can down-rate whoever I want.


You've got some good ideas in there though, I like the idea of ensuring that points used to down-rate a comment are given back to the community, though I'm not sure there's a safe way to do that that's been proposed yet. The Slashdot system is to redistribute the points randomly to average daily readers... though they fuck it up by allowing moderation results to influence this redistribution.
Warning: On Lawn is a documented liar.
[ Parent ]

This is the absolute wrong idea (2.25 / 12) (#53)
by gibichung on Tue Oct 28, 2003 at 08:09:42 PM EST

Why don't we just move the site to a MUD?

Too many people play it like it's a game already, and now you want to keep score?

"No man is above the law and no man is below it; nor do we ask any man's permission when we require him to obey it." -- Theodore Roosevelt

Congratulations (1.44 / 9) (#67)
by pyramid termite on Tue Oct 28, 2003 at 10:11:57 PM EST

You've made the only sensible comment in this long, sad meta-wank.

Remember when we used to talk about something besides ourselves?

On the Internet, anyone can accuse you of being a dog.
[ Parent ]
So, what are you guys still doing here? (1.83 / 6) (#84)
by RobotSlave on Tue Oct 28, 2003 at 11:58:47 PM EST

You especially, termite. You're always going on about how much better usenet is, and how superior the old usenet meow trolls were, and how you're generally a tougher cookie and a bigger player in the world than the rest of us, due to your usenet bona fides.

Are you just sticking around to lord it over us little people who can't handle usenet? I figure there's got to be some reason you're posting at K5 these days, instead of on usenet, but I just don't seem to be able to put my finger on it.

In short: why are you still here?

[ Parent ]

why are you here? (1.60 / 5) (#89)
by clover_kicker on Wed Oct 29, 2003 at 02:21:39 AM EST

Speaking strictly for myself:
I'll tell you this - if I found a usenet group with the same number of interesting folks, I'd stop visiting K5 the same day. K5's only asset is the interesting people, the site's mechanics are vastly inferiour to a usenet reader with a good scorefile ability.

I am the very model of a K5 personality.
I intersperse obscenity with tedious banality.

[ Parent ]
Modded communities (2.37 / 8) (#54)
by Persecuted Telemarketer on Tue Oct 28, 2003 at 08:23:14 PM EST

You mention a few modded communities in your post. Does anyone know of any list of communities which have some sort of moderated / BBS style? I know a few, but I don't know of any vaguely comprehensive list.

The Yankees ARE the Evil Empire, you know. Go Marlins. Go anybody.

normalized ratings and drduck (2.66 / 9) (#55)
by martingale on Tue Oct 28, 2003 at 08:29:25 PM EST

This is just a thought I had reading some comments. It's not meant to be a comprehensive solution to anything, it's just a technical thought I don't know where to put elsewhere.

The comment rating scale on k5 is absolute. Good comments (should) have high numerical values, while bad comments should have low values.

What makes the drduck school of comment moderation an annoyance to everybody is the absolute nature of the scale. Nobody likes to be rated zero, because everyone feels zero means a bad comment. It doesn't matter that everybody is rated zero equally, because everyone is annoyed individually.

However, comment moderation in scoop is used to rank comments only, and ranking doesn't depend on the absolute value scale, only the relative moderation values between comments. You can have a (good,bad) pair of comments with values (2.5,0.6), but the same ranking is also achieved with values (2.5,2.45), or (0.7,0.69).

Now it's interesting that drduck doesn't affect the ranking, only the perceived absolute value of all comments. Provided he rates all comments equally, the ranking is unaffected. This is true of any type of constant rating. The only way to affect ratings is to rate at least two comments differently (which also occurs if you don't rate all comments, ie only rate one comment say).

With this in mind, it suffices to display the ranking rather than the absolute score to solve that problem. The downside is that it is no longer easy to engineer ranking swaps, because it isn't clear how many comment ratings are needed to counteract a given ranking. Maybe that's a good thing.

What's the big deal with Dr. Duck? (none / 1) (#102)
by Kwil on Wed Oct 29, 2003 at 01:12:50 PM EST

I mean, Drduck has rated comments of mine across the scale. Yeah, I've gotten a few zeros, but I've gotten pretty much every other number as well.

And if you take a look, his rating scale is actually pretty easy to understand. It's based on one criteria - is your comment helping anybody with the issue?

You bounce your head off a troll wall and he's going to give you a 0.
You make a pithy one-liner that, while funny, doesn't really add to the understanding of the issue? You're probably going to get a 0 from him.

Oddly, I actually value drduck's ratings as way of telling me, "Hey stupid, you're wasting our time and your time here.. do something else."

That Jesus Christ guy is getting some terrible lag... it took him 3 days to respawn! -NJ CoolBreeze

[ Parent ]
doesn't work for everybody (none / 1) (#107)
by phred on Wed Oct 29, 2003 at 01:27:45 PM EST

I'll remark on drducks ratings of other people, based on that, I disagree. Theres been some spot on, great bottom line lucid posts that get a zero from drduck. And its pretty obvious too, so I'd have to imagine that you're trolling on this one.

[ Parent ]
Why can't we have anonymous story submission? (2.56 / 16) (#56)
by la princesa on Tue Oct 28, 2003 at 08:53:25 PM EST

Nobody's provided a compelling reason why usernames have to be attached to queue submissions.  I'd even go so far as to say that stories should only have the author revealed if the story is voted up.  Hidden stories could be sent to their own unlisted section, by author choice or email request following voting.  It removes the popularity factor from stories, eliminates much of the troll horror element others complain of and forces users at least somewhat to vote according to the actual merits of the story.  I just feel it would be the best way to improve user-submitted content on the site, and I can't see any logical reason it wouldn't be a great idea.  

<qpt> Disprove people? <qpt> What happens when you disprove them? Do they disappear in a flash of logic?
I'll field this one... (2.55 / 9) (#58)
by Talez on Tue Oct 28, 2003 at 09:28:58 PM EST

Nobody's provided a compelling reason why usernames have to be attached to queue submissions.

You mean besides the obvious anonymous crapflooding?

Si in Googlis non est, ergo non est
[ Parent ]

Anonymous? (1.50 / 10) (#62)
by rmg on Tue Oct 28, 2003 at 09:43:54 PM EST

Yes, crapflooders will be protected, temporarily, from the censure of their fellow K5ers. Their wickedness will indeed remain hidden from the eyes of mortal men -- but not from God.

Crapflooders of all stripes spread insidious blasphemies against reason, against humanity, and even against God himself. But do we imagine that they will not get their comeuppance? Nay! For even as the institutions of man and the gentle hand of the editors leaves the crapflooder untouched, the invisible hand of God remains unstaid. As these infidels continue to make a mockery of God's gifts to man -- our rational faculties, our moral decency, our bandwidth -- they forge their chains, the chains they will be forced to bear as they push the many boulders and experience the many tortures in the fiery depths of hell.

The internet may provide a comforting blanket of deception to the crapflooding hordes, it's true, but that blanket can only protect them from the prying eyes of this world. Fear though not, my friend, for God is ever vigilant... ever watchful. I promise you that no wickedness that has ever or will ever befall the queue will go unpunished in this world next.

_____ intellectual tiddlywinks
[ Parent ]

and who's behind... (1.14 / 7) (#63)
by Dirty Sanchez on Tue Oct 28, 2003 at 09:47:00 PM EST

the rise of crapflooding?

you guessed it:

frank stallone.

[ Parent ]

I encourage you (1.00 / 5) (#69)
by rmg on Tue Oct 28, 2003 at 10:15:58 PM EST

To continue the work of the old rmg where the Word of Frank Stallone is concerned. I remind you, though, that when speaking of Frank Stallone that links are both Informative and Interesting.

_____ intellectual tiddlywinks
[ Parent ]

But the incentive is removed in that case. (2.66 / 12) (#70)
by la princesa on Tue Oct 28, 2003 at 10:18:08 PM EST

There's no point if nobody knows anything about it once it's voted down.  It wouldn't even be seen in the unlisted hidden section, were that an option.  It would just be zapped and nothing more.  Crapflooders do seek a certain amount of name recognition.  Anyone interested in just writing a story without the nullo or blah_username crap gets to see how well their ideas and technical style appeal rather than getting voted up or down on popularity or familiarity.  It can't possibly reduce the quality of the articles, and will very likely raise the bar, which is a good thing.  

<qpt> Disprove people? <qpt> What happens when you disprove them? Do they disappear in a flash of logic?
[ Parent ]
*wank* *wank* *wank* (1.44 / 9) (#74)
by Talez on Tue Oct 28, 2003 at 10:37:59 PM EST

You only want anonymous postings because your articles get -1'd on sight by a more than modest number of people. The whole point is so that you can write an article and if it gets posted you pop out of the woodwork and go "HA HA! YOU VOTED UP ONE OF MY ARTICLES!"

If you're considered a troublemaker by the mass populace might I suggest you create a new username instead of trying to push a measure than would do nothing more than encourage mindless crapflooding and fulfil your personal agenda.

Si in Googlis non est, ergo non est
[ Parent ]

It's nothing to do with me. (2.40 / 10) (#75)
by la princesa on Tue Oct 28, 2003 at 10:47:52 PM EST

I've had articles voted up before.  In a couple of instances, they would have made section or FP due to autopost things if the comments within the stories had been more highly rated, which is tangential to the content of the article itself.  I honestly tire of things like 'localroger/rusty/blar_popular_k5_account gets any tripe voted up because of his popularity', or things like '-1, the author's only written 16 comments!  they shouldn't be allowed to have something voted up!', or '-1, the author is a notorious troll!'  All these reasons are irrelevant to the articles themselves.  I've seen far too many wonderful pieces of writing by dozens of authors get voted down for reasons unrelated to the content, and unlike the accounts that dislike me, I'm not so self-centred as to posit a solution that would solely benefit me. I might be voted up in such a system, I might not.  

At present I could be voted up pretty readily.  I simply choose subject matter that has a more unlikely chance at being voted up.  Heck, in my last submission, the vote was evenly split.  That meant 150+ users wanted the thing voted up, which is nearly twice the 95 votes needed to minimally vote something up.  In any case, it really is worth a shot.  Why should username matter good or bad?  The article's the thing, right?  If not, you might try flipping to that paradigm.  

<qpt> Disprove people? <qpt> What happens when you disprove them? Do they disappear in a flash of logic?
[ Parent ]

Clarification: (1.75 / 4) (#77)
by la princesa on Tue Oct 28, 2003 at 10:50:46 PM EST

'in a couple of instances where the article was voted down with a high number of votes for and against)'

<qpt> Disprove people? <qpt> What happens when you disprove them? Do they disappear in a flash of logic?
[ Parent ]
It wouldn't help you (1.06 / 16) (#60)
by Dirty Sanchez on Tue Oct 28, 2003 at 09:36:29 PM EST

anyome can spot your bullshit.

[ Parent ]
Because next time you copy (1.60 / 5) (#94)
by mami on Wed Oct 29, 2003 at 09:41:46 AM EST

from some "great thinkers", I would like to tell the author of the name la princesa, that she/he copied from some "great thinkers".

Somehow I value courageous people, who can take the heat and don't hide behind their anonymity.

If you want to be somebody, I guess you have to become somebody, and I don't think anonymous people are anything near to "somebody".

[ Parent ]

Define anonymous (none / 3) (#112)
by sllort on Wed Oct 29, 2003 at 02:19:03 PM EST

Would it be:
  1. Forced anonymous?: If a story author wants to identify himself/herself, they have to submit the story, then "first post" in their story saying "oooh oooh it's me it's me", or better yet, encoding their username in the story itself? Bad Thing.
  2. Manual Anonymous?: Create a new account, post the story. Hint: we already have this.
  3. Optional Anonymous: On story submission, an option box to post it anonymously. Authors who successfully posted a story would then submit a diary and comments containing "oooh oooh that's my story" messages, and so would a bunch of other people, since there's no way to prove authorship. Also bad.
  4. Optional Delayed Anonymous: On story submission, an option box to post anonymously, which would be removed if the story was successfully posted. So, stories posted in this fashion, if successfully voted through the queue, would lose their anonymity.
Of the four implementations, I favor #4.

Warning: On Lawn is a documented liar.
[ Parent ]
Actually, number 4 is my preference. (none / 3) (#122)
by la princesa on Wed Oct 29, 2003 at 04:06:44 PM EST

While it is probably easiest to make anon submission an option like editing is, as you explained in option #4, i kind of would like rusty to trial run forced anon submission (with author name revealed if article is voted up).  honestly, while there are a lot of assholes, i rather doubt it's an option that would be abused.  i mean, you can already post a story under one account and claim authorship from another, and few enough bother with that.  also, if people had to submit anon by default, you wouldn't get as much crapflood abuse because anon crapflooding is not as entertaining to crapflooders.  the ratio of anon crapfloods to named account crapfloods on slashdot demonstrates this.  i mean, making it optional removes the entire reason i believe k5 should try it.  this site is supposedly about discussion of various isses, not about how popular or unpopular an account is.

Authors shouldn't have to be bothered with their history, or with necessarily having to post n amount of comments to be considered worth voting up.  They should just be able to float their stories and see if k5 wishes to hold them up.  since this site is discussion-focused, there shouldn't be this author-branding that's going on.  it makes both the authors and users lazy and risks the quality of the content submitted to the queue.  maybe if we go anon, some of these interesting stories with scores of 50 and 70 that presently die in the queue would get those extra votes due simply to the discussion value or quality of the article itself.  it's worth a try at least.  

<qpt> Disprove people? <qpt> What happens when you disprove them? Do they disappear in a flash of logic?
[ Parent ]

Ya but but but but (none / 2) (#123)
by sllort on Wed Oct 29, 2003 at 05:24:16 PM EST

My big question here is: how does the story author discuss edits in the edit queue while protecting his/her anonymity (assuming one wants one's anonymity protected)?

Agree still that it should be an option.
Warning: On Lawn is a documented liar.
[ Parent ]

That would involve reforming the edit queue. (none / 1) (#126)
by la princesa on Wed Oct 29, 2003 at 06:26:32 PM EST

It could be done if rusty went ahead and made the edit queue a separate thing from the voting one, instead of this current thing.  Each anon submitter could then edit with those comments tagged with a special id.  Shoot, one could even do things like assign all users an anon-id account that is just for that stuff.  Unique ids for the comments isn't that hard to implement.  you'd have to separate out the edit queue somewhat more than currently though, and perhaps store the comments a little differently, but that's nothing that couldn't be handled pretty readily if one was determined to implement it.  

<qpt> Disprove people? <qpt> What happens when you disprove them? Do they disappear in a flash of logic?
[ Parent ]
Barking up the wrong tree (2.63 / 11) (#61)
by driptray on Tue Oct 28, 2003 at 09:39:41 PM EST

Trying to perfect a centralised moderation system is just so much wasted effort. You'll end up with a complicated system that a significant number of people will still attempt to "game", and everybody else will spend most of their time bitching about.

The simple and foolproof answer is to give up on a centralised answer to "moderation", and to allow individual (and possibly small-group) score files. Let everybody create their own moderation system (ie, score file) if they wish, or allow them to use other people's, or allow them to collaberate in groups of their own choosing.

This is putting power into the hands of the end-users, which seems totally in sync with the underlying architecture of this whole internet thingy. And in this case, that end-user power cannot adversely affect others.

So fuck this dumb search for a moderation utopia. People will always bitch about the moderation - give them a chance to set their own standards.
We brought the disasters. The alcohol. We committed the murders. - Paul Keating

Perfectly correct (2.16 / 6) (#68)
by StephenThompson on Tue Oct 28, 2003 at 10:12:45 PM EST

As long as people I don't trust are able to influence what I read the system is broken. Having per user trust relationships is not difficult to implement; a simple kill file for people I dont trust would go a long way in making the meta topic quiet down.

[ Parent ]
No web community will surpass usenet until ... (1.16 / 6) (#93)
by wobh on Wed Oct 29, 2003 at 09:40:38 AM EST

... it becomes usenet.

That is all.

[ Parent ]
Popularity contests and bathingsuit competitions (2.00 / 8) (#64)
by Fantastic Lad on Tue Oct 28, 2003 at 10:02:59 PM EST

Jeez. I didn't realize that so many people actually cared about those little numbers.

Even on a higher traffic site like Slashdot, it really isn't so hard to scan a hundred or so thread headline comments. On Slashdot, everybody has an equal chance at the podium so long as one has the user settings configured to make all comments visible regardless of their 'score'.

As such, this reduces any ratings system to a simple popularity contest.

Mind you. . . Popularity contests do provide an interesting sample of societal thinking, which I find useful in determining where the polarities of given issues may lie. I like to know how dumb or aware a community is on a given subject, and allowing the peanut gallery to toss in their opinions without having to hack out a formal response each time is useful.

I don't think Slashdot goes about it the right way; they have opted for the jr. highschool level of maturity. I much prefer the K5 system, which results in a much larger sample of opinions. In fact, I like the latest incarnation brought to us by Rusty. The previous 1 to 5 thing didn't work so well because people had different bench-marks for good, bad and ugly. Having only 3 options, and calling the upper and lower end, 'Encourage' and 'Discourage' much reduces that uncertainty factor. Mind you, I do find the 'hide' option a little worrisome. While I understand the advantage for the community to be able to kill off threads by saboteurs, some of the world's greatest thinkers would have also been 'hidden' in their day.


Those little numbers are important (2.50 / 3) (#95)
by straif on Wed Oct 29, 2003 at 10:04:55 AM EST

Reputation is a strong motivating force.  It doesn't fill my belly, but it is no less important.

Last Friday, my slashdot karma went from "positive" to "good".  I am very proud of that.  I've put a lot of effort into my postings.  I don't want to have my voice louder than others, but I do want to contribute in a meaningful way.  My "+3 interesting"'s and "+5 Insightful"'s gave me feedback that my contributions were being noticed.

This encourages me to contribute more, and strive to be as thoughtful and thought-provoking as possible.  The only danger I face is the fear of writing something controversial that may damage my karma.  However, if I write something that I feel is important but unpopular and I do get moderated down, then I consider that positive (if not misguided on the part of the moderators) feedback.  The slashdot system might be improved if they had a "+1 Controversial" moderation option.  Much of what gets moderated down as trolls are merely unpopular opinions.  It can be difficult for the reader to determine what the intent of the author was, but I think experience in electronic fora teaches how to tell the difference.

Perhaps this is shallow of me.  Even if it is, I feel it prods me to make meaningful contributions.  I didn't consider myself more "popular" when my karma went up.  I took it to mean I was contributing to the discussion in a positive way.  I think that knowledge is far more valuable and motivating that popularity.

I don't think I'm the only one.

We are a curious species.  When someone tries to "game" the system it is usually because they are trying to understand it and what holes it has in it.  While any self-described hacker would tell you this is part of the hacker-nature, I think it is more than that: human nature.  As long as the system can deal with the problems that artificially boosted reputation can create it is harmless.  Ultimately, gaming the system will become boring.

[ Parent ]

Slashdoting. . . (none / 1) (#154)
by Fantastic Lad on Fri Nov 07, 2003 at 12:49:03 AM EST

I must admit, when I first began posting to Slashdot many moons ago, I didn't understand their rating system at all, or how Karma worked. When I got a high rating on a comment, I found it exciting much the same way you did. I felt that I was having an impact.

It wasn't long, though, before I realized how the 'game' worked, and I wanted my podium to have the maximum number of phonebooks stacked beneath it, giving me the greatest exposure possible for my 'all important' views (ahem) which I felt it so very neccessary to share with the world.

It wasn't long after that point, after gaining the maximum 'Karma' possible that I realized that it didn't matter anymore. I post whatever the heck I like. I'll blast people, post unpopular views and generally be as abrasive as I feel at the time of posting. Interestingly, I discovered that it doesn't matter what the heck you do, so long as you are coherent, committed and interesting, you will never lose your phonebooks. I've contributed more than 700 or so posts to Slashdot over the last few years, and I generally write comprehensive messages to the world. Rarely one-liners.

The point of my telling this is that all number systems aside, so long as you give the world genuine energy and attention, it will make room for you and listen. You don't have to worry about the numbers. They take care of themselves.

We are a curious species. When someone tries to "game" the system it is usually because they are trying to understand it and what holes it has in it. While any self-described hacker would tell you this is part of the hacker-nature, I think it is more than that: human nature. As long as the system can deal with the problems that artificially boosted reputation can create it is harmless. Ultimately, gaming the system will become boring.

I like this observation.


[ Parent ]

this is a discussion on moderation (none / 1) (#106)
by phred on Wed Oct 29, 2003 at 01:24:32 PM EST

if they're meaningless to you, why are you even posting?

[ Parent ]
Completely uninteresting. (1.00 / 11) (#65)
by Jed Smith on Tue Oct 28, 2003 at 10:04:13 PM EST

I'm tired of meta.
-1. Next.
K5 is dead. Steve Ballmer made the most insightful comment on a story. -- jw32767
Scientific analysis (1.66 / 6) (#86)
by Zombie Uday Hussein on Wed Oct 29, 2003 at 12:17:20 AM EST

Through a state of the art computerized analysis of the poster's writing style, I determine that Jed Smith is one of these 5 users:
  • RobotSlave
  • elenchos
  • trhurler
  • DJ Glock
  • rusty
More datapoints will narrow the list further.

not ZOMBIE turkey. just turkey. maybe a little mayo.
[ Parent ]
Well, Local Roger (1.33 / 15) (#66)
by rmg on Tue Oct 28, 2003 at 10:07:10 PM EST

I have to say that I agree with almost half of what you say here, which is why I did not log in to another account and vote you down twice. That said, I am disappointed with the lack of confused sexuality and talking spaceships in this story.

I did, however, want to comment on one particular aspect of your article. Your idea that discouraging trolls is desirable is really a bit extreme. Certainly, crapflooding ought to be beaten down at all costs, but as I have commented elsewhere, we can safely leave the Lord Almighty to sort out that situation. As for trolling proper, I, amongst other trolling afficianados, find trolls to be a most delightful aspect of any webforum or newsgroup. Indeed, I can hardly imagine a site without a legion of dedicated trolls. It is my belief that if you do too much to discourage trolls, they will turn to vice, i.e. crapflooding. It is very important that we try to limit that state of affairs, which is why I urge all of you trolls and crapflooders to go study geekizoid and Slashdot at -1 as examples of what not to do -- even with a diety as compassionate as ours, we must still work to guide errant sheep back into the flock of the Almighty.

As for specific measures to take to encourage proper trolling, I recommend a troll rating and a crapflood rating. The average K5 user will appeal to his Slashdot intuition and arrive at the conclusion that both must be negative. The reality, of course, will be that the troll rating will register as a 5, while the crapflood will count as a -1. It will be matter of the utmost urgency and greatest importance that true knowledge of the the significance of the ratings be kept strictly confidential. Those who see fit to rate trolls "crapflood" should of course have their accounts closed and their ips banned.

It seems clear that the changes I have outlined above will result in an atmosphere of harmonious trolling with a minimum of crapflooding. In a perfect world, we would have a "whiner" rating for those who accuse others of trolling (patently slander, as trolls are simply an artifact of liberal postmodern though). Unfortunately, we will probably have to settle for the system I have laid out and count ourselves lucky in such a world as this.

_____ intellectual tiddlywinks

Well, RMG (none / 3) (#125)
by mami on Wed Oct 29, 2003 at 06:19:59 PM EST

are trolls nowadays dyslexic or are dyslexics just trolling?

[ Parent ]
I read this a week ago. (none / 1) (#151)
by rmg on Tue Nov 04, 2003 at 09:22:26 PM EST

I have thought about it ever since. I still have no idea what you are talking about. Would you please explain?

_____ intellectual tiddlywinks
[ Parent ]

This is a good framework (2.42 / 7) (#72)
by enfilade on Tue Oct 28, 2003 at 10:24:50 PM EST

Which allows us to apply the large body of economic theory to solving the problem of moderation. It might not be the perfect answer, but you've got a lot of theoretical backup for this system.

However, it amy very well spur on more trolling as people get fake accounts to vote stories FP in order to get a high income. You need to solve the crucial problem of how to stop people who have anonymous proxies making thousands of dupe accounts and subverting the economy to their own ends.

Here are some suggestions.

  1. Story voting should cost mojo.
  2. New accounts start with negative mojo (yeah that's right).
In order to "charge up" the economy it might be worthwhile to dispense equity simply for loading pages (at a low, humanly reasonable, capped rate) to encourage the "silent majority" of lurkers to participate.

No, this would reward obssessive page-reloaders. Instead have a "seed" of moderators that doll out moderation points and keep the economy going. Supply-side economics I suppose.

I think that a seed-based system like Avogadro would be safer to attack. I've got my own theories about a "perfect" moderation system (which I might chance on the queue), and one of the key properties of a good system is that it must resist attack. If there is no central, then trolls can take over the system.

There should be negative interest on mojo to stop inflation, as mojo will only disappear from the system by negative rating; otherwise it just transfers from one person to the other.

And one unrelated point that I've been wanting to make about moderation: There should be as many different ratings (both comment and story) as possible, instead of as few as possible. This is because the bottleneck is the time it takes to make and upload a rating so as much information as possible should be transferred in a single rating.

This does not necessarily mean a slashdot like system, but there should be more choices that actually give information about why a comment is bad or good.

Also, earning an automatic (like the /. +2 bonus) rating is a good thing to have in a system, particularly a negative bonus for trolls and so forth as it allows a system to adapt.

s/avogadro/avogato (2.16 / 6) (#73)
by enfilade on Tue Oct 28, 2003 at 10:29:40 PM EST

damn chemistry!

[ Parent ]
Advogato is one big catch-22 (2.25 / 4) (#99)
by pin0cchio on Wed Oct 29, 2003 at 11:00:16 AM EST

I think that a seed-based system like [Advogato] would be safer to attack.

As I see it, the practical problem with Advogato is that because a new user cannot leave comments at all, a new user essentially has to already have a few contacts on the inside to get in.

[ Parent ]
Hooray (1.50 / 16) (#79)
by shoeboy on Tue Oct 28, 2003 at 11:05:09 PM EST

It's Meta, by LocalRoger and discusses Slashdot! Truly a combination of everything I love best about K5. BEST STORY EVER!

No more trolls!

I'm confused. (1.77 / 9) (#80)
by gzt on Tue Oct 28, 2003 at 11:14:17 PM EST

I shall point out the source of my initial confusion and then comment on the paper as a whole.

Whether you call it Mojo, Karma, "Standing," or something else, all content rating feedback systems have some sort of currency. While there are many different ways of acquiring and spending such capital, nobody seems to have implemented an economy varied enough to be robust. And this is the key to building a system which can be stable in the long term.

Currency is not a good analogy: ratings simply don't serve that function. I also don't know what you mean by robust or stable. From the what you prescribe [and I hope I'm not assuming too much], it seems you mean a system that requires no "outside" intervention to attain the objectives of the operator is stable. Robustness is still a mystery.

However, despite my initial confusion, I was able to comprehend your argument, but I'm still confused. How are these "feedback pathways" supposed to help the site's operator achieve his goals, whatever those are? I see no particular reason for any of your suggestions to do any such thing. Please reveal your hidden assumptions.

I also don't see why multiple "feedback pathways" produce stability in the sense previously defined. Your analogies are inexact and possibly misleading. Please make your argument explicit.

Also, I am left wondering how these changes would feel to the individual user. If ratings became significant, the dynamic of the site would change considerably. Note that the most recent move by rusty was to downplay the significance of ratings! Why do you assume it is a good thing to increase the significance of ratings? Frankly, I'd think the site downright unfriendly if it had even one of these features.

In conclusion: I proudly -1 every article I see, and if it cost me anything to do so, I'd be rather put out.

Wow (2.00 / 4) (#83)
by gzt on Tue Oct 28, 2003 at 11:42:21 PM EST

That was incoherent. I hope you can catch my point: ratings ought to be insignificant. Making them worth something is foolish.

drduck: rate 1 for disagree, 2 for agree, something else if you refuse to play these games.

[ Parent ]

you really got it backwards (none / 2) (#105)
by phred on Wed Oct 29, 2003 at 01:22:11 PM EST

If ratings are insignificant, why bother with a rating system at all, just let anybody at all post anything at all with as many accounts as they want.

[ Parent ]
You know... (none / 2) (#114)
by gzt on Wed Oct 29, 2003 at 02:32:57 PM EST

...it works relatively well. There are occasional indecent incidents which must be dealt with personally by the admins, but is that so bad?

[ Parent ]
If you feel the system works well then ok (none / 3) (#116)
by phred on Wed Oct 29, 2003 at 03:00:35 PM EST

I feel it doesn't. Thats why I find the article interesting, it suggests alternatives.

But thats ok, since I feel k5's moderation doesn't work well, and at the same time don't believe k5's ops actually have the skills to fix it, its a pretty moot point.

[ Parent ]

Alternatives interesting? (none / 2) (#120)
by gzt on Wed Oct 29, 2003 at 03:42:49 PM EST

Bah, I say. Everybody offers alternatives. They're boring, that's why everybody condemns the so-called "meta-wankery". The point of this article was that Mr. Local provided a [muddled and poorly-defined] analogy to promote his alternatives to the current rating system.

[ Parent ]
ok (none / 2) (#131)
by phred on Thu Oct 30, 2003 at 08:22:55 AM EST

you of course know every preference of every user, except you simply excluded my interests, I assume this was just a bookkeeping error, as otherwise, you would have said in your post everybody but me.

So no biggie, I understand errors occur in even the best of the world's mindreading enterprises, so carry on.

[ Parent ]

why scarcity? (3.00 / 8) (#81)
by danharan on Tue Oct 28, 2003 at 11:24:36 PM EST

While "modbombing" should certainly be discouraged, it seems odd to discourage good ratings.

By good, I don't mean just positive ratings. If I have time, I rate every single comment, except those that are nested. 0 is for obnoxious trolls, but I encourage ratings that are interesting even if I don't agree with them.

Since we rely on large numbers of people to make ratings meaningful, I submit that this is useful and should be encouraged.

Some of Enfilade's suggestions for starting with negative mojo, negative inflation and an advogato rating system could work well with this.

The way I think it should work* (1.75 / 8) (#85)
by bhtooefr on Wed Oct 29, 2003 at 12:08:53 AM EST

* after ten minutes using K5, and almost 1 year on /.

I think that /. and K5 both have good ideas here. I like the K5 mod system (except it does have it's eccentricities), in that it allows you to mod without the /. modpoint server having to pick your username. However, I think that it should be supplemented by a karma system and RESTRICTED modpoints. The restriction could be as simple as a max of 10 modpoints blown per thread or story. Now, it's not about having modpoints, it's about their effectiveness. Let's say at new account creation, you've got a karma of 0.00. This is the multiplier on your modpoints (however, if your karma were negative, the modpoints would count at 0.00). Max karma is 5.00. A STORY would have a *2 multiplier in addition to the mod points. Someone would have to work out the scale who knows K5 very well, but it could go like this:

New Acct gets modded up by new acct 2: no effect
New Acct gets modded up by someone at 5.00 karma: 0.50 added to karma
New Acct now has a karma of 0.50, and then mods down that same person: 4.995
The decimal places would get REALLY long REALLY quick, but it would give a very reliable and powerful mod system.
I'm too lazy to put my sig here. Go to /., fool.

you got it wrong (2.37 / 8) (#88)
by cbraga on Wed Oct 29, 2003 at 12:43:48 AM EST

the idea here is that everyone moderates and gives their vote on the merit of each comment. then they average each other out and the result should mean something.

ESC[78;89;13p ESC[110;121;13p
[ Parent ]
no (2.25 / 4) (#104)
by phred on Wed Oct 29, 2003 at 01:20:59 PM EST

not everybody will moderate effectively. Thusly, only those who earn reputation should be allowed to effectively moderate.

[ Parent ]
my system (none / 1) (#141)
by bhtooefr on Thu Oct 30, 2003 at 06:06:39 PM EST

would still allow anyone to moderate. It would just allow people with higher karma to moderate more. Basically, would you listen to someone who hadn't earned respect over someone who had earned top respect, even if they were making the same point? Would you let Joe Blow Linux User make Linus' keynote? Now, if JBLU acheives OSS-God status, you might do a parallel keynote...
I'm too lazy to put my sig here. Go to /., fool.
[ Parent ]
less competitive (2.57 / 7) (#98)
by dimaq on Wed Oct 29, 2003 at 10:38:54 AM EST

I think what you're discussing or proposing, would mean a system where users compete for 'goodness' points. I don't think it's a good practice, because this is, mainly, a discussion site, not a competition.

(that is a well-motivated troll can post bottom-licking comments just to get credit and then use it to spam everyone or something)

the "goodness" you speak of (none / 3) (#103)
by phred on Wed Oct 29, 2003 at 01:20:12 PM EST

is really just reputation. This alone solves much, ie., you can clone accounts all day, but it takes effort to build a reputation. If you look at what this means, ie., take a real bottom line attitude, you'd see this differently. I like discussion, but lets face it, much discussion on here is nonsense, and much of that can be blamed on the broken moderation.

[ Parent ]
point being (none / 1) (#149)
by dimaq on Tue Nov 04, 2003 at 06:15:16 AM EST

that for a "place" where a reasonably minded person spends 15 minutes a day tops, there should not be such a requirement as build more trust than others before you are accepted.

(and no you could not start of with some trust, cause that would be the avenue for trolls)

[ Parent ]

K5 ratings would be much improved (2.69 / 13) (#100)
by wurp on Wed Oct 29, 2003 at 11:05:24 AM EST

In my opinion, K5 ratings would be much improved if we could ignore other users ratings.  I want an ignore list, then when we get a drduck I can just put him on my ignore list so he doesn't fubar the ratings I see on posts without me knowing it.

Such people would weed themselves out of most users' field of view.
Buy my stuff

New K5 ratings system (2.16 / 6) (#101)
by wurp on Wed Oct 29, 2003 at 11:08:13 AM EST

It seems that most ratings are either 3 or 0.  People are not using 1 the way they are intended.  This problem seems somehow to be much worse than it was with the old 0-5 rating system for some reason.
Buy my stuff
customized ratings and rewards (2.50 / 4) (#111)
by hildaur on Wed Oct 29, 2003 at 02:10:09 PM EST

One could, perhaps, reward honest rating by giving the user comments they like.

First, let every user rate everything they want. After the user has rated some sufficiently large number of posts (this number could take some tuneing), the score for any given post shown to that user can be calculated based on the users tastes: when the ratings are averaged to produce a final score, the weight of any given moderator can be determined by the correlation between that moderators past ratings and those of the user to which the score will be shown.

If this is done in a forum where one can easily select only highly rated posts, the rewards and punishments fall out naturally; you get to see the posts you want to, and avoid those you don't.

The only problem is keeping newbies long enough for the benefits to become apparent. Perhaps the median ratings of posts left unrated by a user (such as those from before the user arrived) could get some small weight in the calculation of the correlations.


Processing power (2.00 / 4) (#113)
by Gord ca on Wed Oct 29, 2003 at 02:32:13 PM EST

You do realize this would take an inordinate amount of computing power to do such a calculation for each comment rating for each user. Rusty doesn't have unlimited computational power to run this site, you know.
Or maybe it wouldn't, I can't be sure. This moderation system is sounds like you could write a PhD thesis on it.

If I'm attacking your idea, it's probably because I like it
[ Parent ]
perhaps, perhaps not (2.00 / 4) (#115)
by hildaur on Wed Oct 29, 2003 at 02:59:17 PM EST

If there are lots of users, I'm sure the full blown method would require a fair amount of compute power. Then again, modern cpu's are a lot faster than one might think. I don't have a real good feel for how reasonable this is either.

A lot of short cuts could be taken to reduce compute power required without reducing effectiveness much. Some things worth trying would include a binary (worth reading/not worth reading) rating system, only keeping the most recent x ratings made my any given user, and only cross-correlating until you reach some minimum number of other moderators that meet some agreement threshhold. The last is particularly important if there are lots of users. I'm sure there are other things you can do.


[ Parent ]

Well when I last proposed your system (none / 3) (#117)
by sllort on Wed Oct 29, 2003 at 03:22:00 PM EST

I suggested it be for subscribers only.


Which solves the computing power problem.
Warning: On Lawn is a documented liar.
[ Parent ]

Subscribers only? (none / 2) (#136)
by RadiantMatrix on Thu Oct 30, 2003 at 11:26:12 AM EST

I'm not sure offering this as a subscription-only service would entirely solve the computing power problem.

With an adaptive moderation system such as proposed being available to subscribers, more people would (theoretically) subscribe.  What happens when subscriber mass reaches a critical point, and the cost of computing power required surpasses the income from subscriptions?

I don't like spam - Parent ]

Well, (none / 2) (#138)
by sllort on Thu Oct 30, 2003 at 12:59:36 PM EST

you take their money, and you buy enough computing power to do it.

Or, you price your service incorrectly, fail miserably, and go out of business.

Maybe he should get an MBA.
Warning: On Lawn is a documented liar.
[ Parent ]

My question... (none / 2) (#142)
by RadiantMatrix on Fri Oct 31, 2003 at 01:02:54 AM EST

This seems like a non-linear problem to me: i.e. the computing power required does not grow in a linear fashion in relation to number of subscribers.  So, is it even possible to both (a)price such a service so that it remains profitable and (b)remain below a price point where people will still want to pay for the service?

I don't like spam - Parent ]

Not dependent on number of users... (none / 1) (#148)
by hildaur on Mon Nov 03, 2003 at 03:16:39 PM EST

The speed of this sort of approach does not need to depend directly on the number of users, but only the number of users who have rated comments. For example, when a user is to be shown ratings for any given comment, the server could look through the list of people who had rated that comment, checking the correlation between each rater and the user. Once it reaches 5 or so raters who match the user well, it could stop and just use those 5. You could also do things like cache the correlations.

In the extreme case of user who only correlates well with 10% of raters, you would not typically have to calculate more correlations than 50 per rated item. If the correlations were cached, it would be far, far less.

The most I think about it, the more skeptical I become of the scheme really needing a lot of compute power.

This has the peculiar effect of slowing the system down for users with unusual tastes.

[ Parent ]

Not really... (none / 1) (#144)
by Gysh on Fri Oct 31, 2003 at 02:11:33 AM EST

Not really - it would depend on how it was done, but (assuming I read the parent comment right) it probably wouldn't take that much comptuing power. For example, if the number was only calculated whenever it was needed on a single-user basis, or if you only calculated it for every user at, say, 1am each day, it wouldn't have an adverse effect on the site.

[ Parent ]
Am I welcome? (2.66 / 9) (#132)
by mcherm on Thu Oct 30, 2003 at 08:59:28 AM EST

I read K5 a lot more than I post. Sure, I post occasionally (and have submitted stories in the past), but I figure that my biggest contribution to the signal-to-noise ration is to simply keep my mouth shut when I don't have much to say. However, I _do_ try to contribute to the site, by meticulously rating all of the comments I read. I try to give ratings intelligently, and hope that this will help guide later readers to the more insightful posts.

In your system, my approach would be too expensive. Am I welcome?

-- Michael Chermside

This was exactly why... (2.40 / 5) (#134)
by localroger on Thu Oct 30, 2003 at 10:29:13 AM EST

...I proposed passing out equity for reading. This should of course be subject to sanity checks and probably capped fairly low to keep people from using bots to pump up their equity, but it's also a way to redistribute the currency that is lost from the system when you spend equity to downrate something.

What will people of the future think of us? Will they say, as Roger Williams said of some of the Massachusetts Indians, that we were wolves with the min
[ Parent ]
currency creation/destruction (none / 3) (#146)
by MikeWarren on Fri Oct 31, 2003 at 06:42:53 PM EST

what's not clear in your story is how exactly currency supply would be regulated. in economies like ours, most money is created via debt (i.e. when a bank loans money, it is creating currency; when you pay it off, the currency is destroyed).

It seems implicit that creating a new account creates currency. Most of the "spending" activites seem to be transfers (i.e. if I rate your story up using 2 credits, do you get those two credits?) but they could be currency-destruction (i.e. the two credits vanish).

It would seem to make sense to make the currency supply related to the amount of content and users; more content, more currency (more users, more currency).

A debt-based approach might look like this: in order to post a story, you take out a "loan" (i.e it costs currency to post) which -- if it's high quality -- you will expect to get back as people "pay" for your post. The site (i.e. bank) could then regulate the amount of content easily: make it cheaper to post if participation is low, make it more expensive as participation increases.

Combined with trust, such a system could encourage people to keep a single account by giving monthly (/daily/weekly) stipends of currency depending on their trust (i.e. highly-trusted users get lots of currency every month). Hence, the richest users will either be voluminous posters of high-quality material or highly-trusted (or both). Presumably, these are the people best qualified to rate comments and posts and so it is not a problem that they are "rich".

-- mike warren
[ Parent ]
Doesn't work. (none / 3) (#150)
by mcherm on Tue Nov 04, 2003 at 10:35:25 AM EST

localroger writes:
> This was exactly why...
> ...I proposed passing out equity for reading.

But either (1) the equity for reading is enough to rate all or nearly all of the comments read (so long as very high or very low scores are given only rarely), or (2) it isn't.

If (1), then no one will ever lack for equity.
If (2), then I couldn't afford to rate things.

-- Michael Chermside
[ Parent ]

trust (none / 1) (#145)
by MikeWarren on Fri Oct 31, 2003 at 06:31:39 PM EST

I think trust is a useful concept for rating systems, especially for encouraging users to keep a single account; if users can give trust-ratings to each other, many complicated (or simple) uses of the resulting trust-network can help users sort out quality content (and, remember, "quality" is subjective, so posts that I like aren't always going to be posts that you like).

So, for example, maybe I can turn my preferences to highlight posts from authors I trust (simple). This can extend arbitrarily far through a trust network; if someone I trust 50% trusts someone else 50%, does that mean I should trust that person 25%?

It might also make sense to trust content itself; this trust could filter (degraded) out to authors and perhaps even commenters (i.e. if someone makes a highly-rated post on highly-trusted content, should that post get an implicit trust increase?)

Anyway, some interesting ideas. Viewing ratings as economic activity seems quite plausible, but I think that filtering is somewhat different. The above trust ideas are mostly filtering (e.g. showing only trusted authors' posts) and wouldn't (IMO) work very well with any of the rating systems described. It seems that "rating" is a global (i.e. site-wide) activity while "trust" is personal; the latter could be used to come up with values for the former, of course...

-- mike warren
Parallel (none / 1) (#147)
by Scrymarch on Sun Nov 02, 2003 at 08:44:49 AM EST

Well, for a start, I don't think k5 does fit this system, because Mojo has been disabled.  That was the equity analogue.  The new state of k5 has no explicitly recorded metric for reputation.  I think K5 still does have a moderation economy, but it's more like the parallel and implicit local currencies Bernard Lietater likes to talk about.  There's a gift economy in fives (threes), and a scarcity economy in zeroes.

When you rate things 3 it's as part of a happy go lucky mutual appreciation society.  The diaries are a good example of this - it's still common practice to rate people up as a way of saying - "hey, thanks for dropping in!".  They promote more high ratings in return, at no cost except time.  The zeroes, however, are a chore, and annoy people, provoking retaliation.  To much retaliation and you used to lose your ability to give zeroes - scarcity.  As a result they tend to be hoarded more.

The changes - thresholds, plus zeroing for everyone, has limited this effect.

Economy growth and inflation (none / 2) (#152)
by cribeiro on Wed Nov 05, 2003 at 05:35:24 PM EST

Now that we are talking about currencies and monetary systems, there is need to think about other important economic issues such as economic growth and inflation. What follows is a rather simplistic attempt to outline some of the issues. I'm not an economist by training, so there is a (probably big) chance that I'll get it wrong; I'm grateful for any help or guidance on the subject.

The system outlined above does not take into account the actual effects of economic growth. To register new customers, the system must issue currency. This is one of the base causes for inflation (it's a simplistic model, but anyway, we're just starting to work it out it now). That is, my currency is worth a little less, because there is more money floating around the system.

One of the ways to counterbalance the effects of inflation is with economy growth. Classical economic theory states that some inflation is needed as part of the growth process. So, how can the K5 economy grow? There must be some way to encourage people to think about the return on investment. It works as follows: when I rate a comment or story, I'm spending part of my currency. I usually shell out money for one of two reasons: either as an expense, or as an investment.

The model proposed treats rating as an expense. Well, I don't see it working this way. Some measurement of investment has to be part of the mix. But beware - too much, and the system will be open to attack again, as people will modbomb to generate money, but no value for the community. It has to follow a curve of diminishing returns. I have two proposals for now, but others will surely be put forward:

  • Give back some of the investment to the rater. It works like a simple lease. If a story that I rate gets a good rating, part of the money returns to me. It has to be a very small amount, just something to foster growth, but not to enable attacks.
  • Make it work like a stock trading system. People would be able to buy or sell stakes on stories at will. But stock exchanges are very complex beasts, and I would not even get close to it. But it may be a way to study and solve the problem.
What about taxes? It may turn out to be a good way to remove excess money from the system or to redistribute it. It may not be needed in the K5 system, where there is little incentive to have your money sleeping somewhere; inflation will crunch it, sooner or later, as the actual amount of currency in the system increases.

And, last of all, one of the most visible effects of inflation... the visible devaluation of the currency. It may be funny to see that a good comment needs a +3 rating today, but the same comment a few years from now it may need a +7, or +20, to be worth reading. That's inflation at work. In a capped system like today's (where a brilliant comment is always a +3 at any given time), I'm not sure on how could inflation be modelled. But it's worth exploring it.

economy and reputation systems (none / 0) (#156)
by typaldos on Thu Feb 12, 2004 at 07:55:50 PM EST

Actually, there is no need to define the "perfect" reputation system. If the process is done at the level of business rules, these rules can evolve as the site evolves. In fact, each member of the site could have his/her own reputation rules and apply them to the site to get a personal view. For example, a member might want to rate the comments of all of his/her trusted colleagues higher than the comments of others.

At my last company, RealCommunities, we actually build a generalized reputation system driven by a business rules engine. The engine accepted inputs that were generated by members, by applications, external events, etc. The site owner determined the actual business rules and could do this at a "channel" level e.g. points for doing something in Channel A didn't necessarily affect the reputation of the member in Channel B. As a simple example, the eBay reputation system provides a 1,0 or -1 score for a transaction regardless of what the member sold. Obviously building up a good reputation selling beanie babies is not the same as a good reputation in selling real estate.


See also: http://www.profguilds.com http://www.resumeblogs.com http://www.typaldos.blogspot.com

Notes Toward a Moderation Economy | 155 comments (140 topical, 15 editorial, 1 hidden)
Display: Sort:


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!