I respectfully disagree with this proposal*. A lot of the reasoning for such a system appears to deal with wasting votes, which is going to be the only bullet point i comment on.

Before you downvote a node, go to Worst Nodes and see if that node is already there. You can see the node rep on that page without having to vote for it. Also, before you downvote a node, ask yourself "Wouldn't this vote be better spent as a ++ on a node deserving of such?" Trolls don't eat votes, but they do eat disk space and bandwidth (which i don't think this proposal will solve as people still feed trolls even if they do realize it). Isn't our Consideration Process fully capable of labeling troll-ish posts?

L-LL-L--L-LL-L--L-LL-L--
-R--R-RR-R--R-RR-R--R-RR
B--B--B--B--B--B--B--B--
H---H---H---H---H---H---
(the triplet paradiddle with high-hat)

My objection to your suggestion is that it's not workable. By that I mean it's not being done, and you've proposed no change to cause it to happen. Sure, people could do it, and probably some do, but enough of them aren't that trolls continue to get big responses.

One reason your suggestion isn't being followed is that it's cumbersome. It's also non-obvious. You're asking people to navigate (without providing a nearby link) to another page to take an action on the current one. It's simply not going to catch on, particularly among the less-experienced.

We could provide a link to Worst Nodes on every post that has made the list, but that would be a badge, which I'm trying to avoid. It would also be kind of harsh to the people whose posts are at -1 or -2.

Isn't our Consideration Process fully capable of labeling troll-ish posts?

I don't know what you're referring to. There's no labeling option in consideration.

---

Caution: Contents may have been coded under pressure.

(Disclaimer: this is all personal opinion if it is not apparently so.)

By that I mean it's not being done, and you've proposed no change to cause it to happen.
But it is being done (even if i am the only one doing it -- it is being done). And the fact that i have proposed nothing to change it is my point: i see this whole issue as a solution looking for a problem.

... but enough of them aren't that trolls continue to get big responses.
And i fully believe that they will continue to get big responses even if you label the post as potentially trollish. But that's just my opinion.

You're asking people to navigate (without providing a nearby link) to another page to take an action on the current one.
If our audience was one of typical non-programming computer users who barely know how to read a manual and fix basic problems on their computers, then i would agree with this. But our audience is not. The core audience is experienced Perl programmers. Thanks to tabbed browsing, this is much easier than you say it is not. Most of the less-experienced users you speak of are here to ask questions, not answer them. It's the experienced users who will be answering the questions, giving feedback, etc. It is they who will be potentionally replying to trollish posts.

I don't know what you're referring to. There's no labeling option in consideration.
But there is. When you consider a node you have the option to give the reason why. That text appears on the content for the reaped node. See? You already have a way to label a post as trollish and as an added benefit, everyone else with consideration power (that is, those who are not the "less-experienced") can cast their vote as to whether or not they agree that it is trollish.

Again ... take a look at who is doing the responding. We are mature, experienced Perl programmers. If we take the bait and feed a troll, we know what the consequences will be. The consideration process was put into place years ago to handle this problem, and i fully believe that it is handling the problem as best as anything can. If a human cannot always determine 100% of the time that a post is truly a trollish post, how can a program do the same? Some people just can't get enough of that negative attention, and no system is going to stop the occasional trollish node from being posted.

jeffa

L-LL-L--L-LL-L--L-LL-L--
-R--R-RR-R--R-RR-R--R-RR
B--B--B--B--B--B--B--B--
H---H---H---H---H---H---
(the triplet paradiddle with high-hat)

Dear Roy Johnson:

As usual, a very well thought post (++). Just a quick comment...

We could provide a link to Worst Nodes on every post that has made the list, but that would be a badge, which I'm trying to avoid.

But the rep, in itself, is a kind of badge. Although you see it after your vote is cast, it may very well influence your criteria for voting nodes for the same monk that come later.

It would also be kind of harsh to the people whose posts are at -1 or -2.

I agree, although these either won't end up in worst nodes (at least, for a long time).

Best regards

-lem, but some call me fokat

People see a horrendous post and think it's important to downvote it, not knowing that it's already got plenty of downvotes. This is significant to those who are on a vote budget.

I don't understand why this should be true. Isn't everyone on a vote budget? Oughtn't people vote on their perceptions of the merit of a post without regard for information gleaned elsewhere?

Update: tr/m/n/ where appropriate (thanks, radiantmatrix!).

It has nothing to do with vote budget, but I've often wished that I could see the various votes on a node prior to reading it and/or voting on it, and that this feature was configurable in user settings. See votes all the time, see votes only after voting, never see votes. Yes, I know that I could see some votes this way so long as they made it to Best/Worst nodes, but what about those that didn't?

Whether or not other votes impact the way I vote, if I choose to, is beside the point. It is more a matter of content filtering. Sometimes the number of votes tell you something that the title of the node does not. Isn't this effectively what occurs when you are browsing "Best/Worst nodes?" You're viewing nodes that people felt positive enough to vote for or against. Perhaps they agreed, Perhaps they found it useful, or (unfortunately) maybe they just like who posted it...

It could affect whether or not I choose to read a long post, and it may give me an additional way to judge the reliability of a response. Not a definitive way, but one nonetheless. There are lots of nodes, and I don't have time to read them all. So I figure a "rough" ranking is sometimes better than no ranking (A broken watch is exactly right twice a day, and one that is 10 minutes off never is, yet which is more useful?). Sometimes I'll read a node solely based on the person who posted it, or even because of a person who responded.

I find that I sometimes will use this approach when browsing through user-submitted book/movie reviews on other sites. Sometimes, I find myself wanting to focusing on the extreme good, and the extreme bad reviews, and skip the middle-of-the-road comments. I also see it similar to viewing survey results without ever taking part in the survey.

I don't think I get any more value from seeing this information once I've voted, than before. It just tells me how others felt about the same node. At that point, generally speaking, I don't really care how others voted. I've already read it, formed an opinion, and cared enough to vote. I will likely be more interested in what other people had to say rather than vote at that point.

---

Update: I've thought about this posting several times since writing it, and I haven't read anything since that's changed my opinion, but being a bit critical, I think I mispresented my position and may have given the impression that the reputation was sufficiently useful to use by itself as a sole indicator on the value of a node. While I believe reputation can have value, I think that value is lessened when taken in isolation of other criteria.

I primarily browse Perlmonks through Recently Active Threads, and use whatever information is available to determine whether or not I'm going to look at a node, much less vote on it. While this may be slightly unfair to judge a node by it's title, author, number of responses, and identity of those who respond, I don't really care.

I have a finite amount of time in which to browse the site, and that doesn't allow my the luxury of reading everything. I view the reputation as one more potential criteria that could be useful.

An isolated reputation alone tells you one story, and not the whole story, however a reputation taken in the context of a particular thread may have value. Nodes in topics that attract a lot of interest (and often responses) tend to have higher reputations than nodes in niche topics, so comparing nodes in different threads seems useless to me. Witin the context of a thread however, I find that it can serve as a general indicator of perceived value. How did individuals who were interested in the subject matter rate the node... that sort of thing.

Just a few parting thoughts on this topic, since I don't expect to author many more (if any) nodes on this subject again.

One node of mine has a reputation of 6 right now. Another has a reputation of 14. One has a reputation of 39 and another 81.

What can you learn from that?

By focusing on the particular value of a node's reputation, you can answer only one question: what is the current value of upvotes minus downvotes for this particular node? You cannot answer several other statistically relevant questions such as:

• What is the total number of people who voted on this node?
• What is the total number of people who read this node?
• What is the total number of people who would have voted on this node if they had sufficient votes to do so?
• How many votes did this node receive in comparison to other nodes today?
• How many votes will this node receive in the future?
• What is the average daily distribution of votes by time?
• What is the current reputation of this node in comparison to all other nodes in this thread, of the day, and in the system altogether?
• What is the current maximum or minimum possible value for this node?

There are likely many other important questions. Looking at even that minimal list, it's easy for me to say that the single numeric reputation of any single node is exceedingly useless. You're welcome to intuit some shade of meaning from that number, but that single number is admirably short of any sort of context that can put it into a statistically relevant and analyzable setting.

Isn't everyone on a vote budget?

Some budgets are tighter than others.

Oughtn't people vote on their perceptions of the merit of a post without regard for information gleamed elsewhere?

That encourages trolls, as I mentioned.

---

Caution: Contents may have been coded under pressure.

Isn't everyone on a vote budget?

I'm not. If I run out of votes for the day, then hey, I run out of votes for the day. FWIW, this has only happened once since I reached a high enough level to get a picture on my home node. But no, I don't budget my votes. I just spend them. If I see a node that I think is notably good or bad, I just up or down vote it (respectively). But I get so *many* votes, how could I use them all up? I'd have to read Perlmonks eight hours a day; I'd either never get any work done, or I'd never get any sleep.

I'd have to read Perlmonks eight hours a day; I'd either never get any work done, or I'd never get any sleep.

So what's your point?
:-)

Do not rebuke them with harsh words ... but rather lead them gently - with URLs - so that they may learn wisdom.

I've been here almost 5 years now, and I'm still amazed at how much time and attention is spent on voting and XP.

++ for a thoughtful idea, but I agree with some others that this probably wouldn't be very effective in avoiding trolls. I am somewhat vulnerable to trolls - I tend to respond to them, argue with them, and ultimately waste my time with them. I almost never downvote them - I prefer to upvote worthwhile posts, rather than to downvote useless ones. I seldom "waste" a vote. What I do waste is time - so unless you could magically make troll posts vanish, chances are I'm going to get sucked into them and argue - even if the post has a flashing icon saying, "TROLL".

Another objection I have is simply that posts often get downvoted because they make an unpopular - but not unreasonable - claim. What the votes measure is not the worth of a post, but simply its popularity. Some well-reasoned, respectfully worded posts have been downvoted simply because they discussed something that was unpopular. Chances are the OP's post itself is among those. I haven't yet voted for it, but I wouldn't be shocked to see a significant number of downvotes.

A troll alert is likely to incorrectly identify at least some of these reasonable, but unpopular, posts. I think this would hurt Perl Monks in the long run.

Finally, it seems that Anonymous Monk's posts don't really get all that much downvoting. Limiting the maximum negative rep to -9 doesn't seem like it would accomplish much. Most of the really bad nodes are by known Monks.

Some well-reasoned, respectfully worded posts have been downvoted simply because they discussed something that was unpopular.

Oh boy, how I wish more monks made that distinction. ++

---

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

Lingua non convalesco, consenesco et abolesco. -- Rule 1 has a caveat! -- Who broke the cabal?

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.

posts often get downvoted because they make an unpopular - but not unreasonable - claim ... Some well-reasoned, respectfully worded posts have been downvoted simply because they discussed something that was unpopular

Really? Can you provide some examples? I'm not just being rhetorical, I really can't think of any instances of this happening, and I think some examples would be helpful to demonstrate your point.

+10 ()++  -4 ()--  = +6 ()+0
[download]

All systems work best with full information, save for those systems that wish to deny the ability of others to do things, like governments and corporations. This isn't to say that governments and corporations don't do important things, but they do them based on denying capabilities to the individual vs. enabling the individual to do things. A lesson can be drawn from that.

"All systems work best with full information, save for those systems that wish to deny the ability of others to...

Actually, dragonchild, though I agree with your observation as a general truth, the current voting system here has the merit of being, IMO, one of several exceptions -- perhaps even the poverbial "exception that proves the rule."

Hiding vote totals until a Monk makes her/his own judgement (except when they're available in best or worst nodes) avoids -- to some extent, the "piling on" or "crowd psychology" phenomonon.

So, when I have votes again, I shall upvote Roy_Johnson's post, not only for its thoughtful presentation and straightforward acknowledgment of many of the negatives voiced on the CB, but also because I see some value for those new to perl and the Monastery in offering a warning that the community concensus on a node is that it's somehow "bad" -- When code is involved, seems to me most often to be the case when that code is ill-considered or even 'flat out wrong.'

Contrarily and sadly, that's clearly not always the case for nodes expressing opinions... be those rehashes of the editor crusades or preferences among modules.

And the preceeding paragraph is a reason my upvote will be for the node; not necessarily for its implementation -- at least, not until we've considered its benefits, costs and debits, and perhaps synthesized an improved proposal.

We already have this: Display of Node Reputation.

The better solution would be to display something like "This post has received N upvotes and M downvotes".

This is equivalent to saying "This post has reputation X". Right now, this doesn't exist (and, IMO, for good reason). I think to implement this would be to cause a bevy of popularity voting. I'd rather see each post stand on its own merits. Some indication that a given post is a troll would be fine, but giving out specific numbers is (again, IMO) not the best.

thor

Feel the white light, the light within
Be your own disciple, fan the sparks of will
For all of us waiting, your kingdom will come

I think to implement this would be to cause a bevy of popularity voting.

Both you and ww made this basic statement, and maybe it's the sociologist in me, but the premise interests me. I'm still relatively new and haven't really paid too much attention to nodes that had anything to do with experience, except for this one. I guess I've missed out on the various arguments you and others have made in the past. Please correct me if I've got this wrong:

I see that others have voted a certain way on a node, so I'm supposed to feel like I should automatically vote for it, too? (Basically the herd mentality) I suppose I can see this possibility, on the other hand, I think it's equally possible that a node was upvoted because a number of individuals thought it had merit. The fact that it received positive recognition attracted others to the node, who might not normally have looked at it, and also determine for themselves that the node, and perhaps others responding to it were worthy enough of an upvote. This seems like a fairly natural effect - more a result of seeing what all the fuss is about than being a sheep. It's a difficult point to prove either way.

What amuses me is that my first inclination when I see a massively upvoted node is to try and determine for myself what's so great about it. I might even judge it far more critically than another less popular node, figuring that it must be really impressive if so many people voted on it. I might hold it to a higher standard.

Still, even if the premise is true, aren't there already ways in which this could happen on this site? For example, let's say for a given node, I see that there were a lot of responses, depending on whatever my personal definition of "a lot" is. I would guess that this attracts attention as surely as upvotes would. Also, don't the "best/worst nodes" encourage the same mentality?

Node Tension

I'm not claiming it's a total match to the topic, but it does discuss some of the dynamic.

It generated a moderate amount of interest at the time, but what level of interest engenders real change around here I'm not sure -- it's quite possible that today is a different climate from when I first wrote it.

But some of the ideas at least overlap with what you're suggesting.

Cheers,
Matt

You know, that idea is something that's been tugging at my brain - couldn't really figure out what it was, but you said it so well. There really is tension over some posts - such as, the OP's in this thread. I'm not sure whether keeping track of tension would be worthwhile (or how much work it would create for the gods), but it's an interesting concept.

Interesting node, Matt, particularly the notion of reporting on tension in threads. For kicks, I threw together a little code to see how well the binomial standard deviation would work as a measure of an individual post's tension.
Read more... (752 Bytes)
The results didn't appeal to me, because a lot of votes in one direction and just a few in the other can make for a comparatively high result. So using the stddev of reps in a thread as a measure of tension in the thread probably wouldn't appeal, either.

Possibly counter-intuitively, it is the nodes of low tension (and low rep) that I would like to have warnings for, while I think that threads of high tension are most likely to include trolls.

---

Caution: Contents may have been coded under pressure.

User setting: "Hide" nodes with rep < -1-$NORM/2.

Currently, the only places we "hide" nodes are Newest Nodes, Recently Active Threads, and Search Results. Those would be fine, but hiding replies seems more to the point.

I'm against labeling nodes based on reputation (except numerically). But I don't currently foresee many problems from adding the feature I described. I don't see a pressing need for it either.

Many replies seem to have missed the point that the proposal is for a heads up ahead of voting to avoid feeding trolls and to avoid wasting a vote on a node that has already received a hamering.

I guess for symetry you could set a "Billy Goat" (just to pick a name at random) threshold for highly rated nodes.

There are good reasons for not showing the rep for a node before you have voted on it. OP is not proposing a change to that. It is a change that you could adopt or not as you wish. It will have little effect on the normal operation of PerlMonks, and can lead to a large reduction in Troll like activity. Where's the down side?

P.S. I presume all the down voters in this thead are Trolls themselves :-)

I haven't been that active here the last two years due to various outside things, but have trolls become that large a problem here?

I still think that there's no such thing as a wasted vote.

-Lee

perl digital dash (in progress)

I found your proposal interesting as I have been thinking of something similar. What i was thinking is that any nodes underneath a reaped node become "dead nodes". That is they can't be replied to, they can't be voted on and they wouldn't show up in worst nodes or in any of the node lists like NN or RAT or the like.

The point would be to cool heated threads, and reduce the "snowball" effect of a troll and those who feed them.

Trolls eat votes. People see a horrendous post and think it's important to downvote it, not knowing that it's already got plenty of downvotes. This is significant to those who are on a vote budget.

As an alternative, I would propose that posts by Anonymous Monk have a minimum reputation limit of -9. Once a post has reached that limit, the downvote option is disabled. That would act as an indicator and also eliminate the negative-rep badge incentive for Anonymous trolls.

I also think you're solving the wrong problem. Trolls are not the problem. Troll droppings are: the replies. And considering there are always people eager to reply, the damage will be done before enough downvotes have been collected. For a flag to warn people about the existance of a troll, you need the flag in place before people can reply. So you might want to disable making replies before a node is at least X hours old, for some X. Which I don't think is what people really want.

Doesn't the same hold for good posts?

No, because upvoting isn't being baited. People downvote to discourage posts like the one being voted one, but trolls aspire to acquire those downvotes. So the downvotes are wasted (are, in fact, counterproductive). Upvotes do exactly what they're intended to do, so they're not wasted.

real trolls care about responses

As I said, I expect responses to be reduced when people know that the post has already been widely dismissed.

Just go to the sign-on page, enter an email address of 'troll@mailinator.org', collect your password, and you're off.

That gives the monks more information, so I see that as a benefit, not a tactic that completely negates my proposal. People notice new names and see that it's their first post or remember that they have a history of bad posts. That's why most trolls use Anonymous Monk.

Trolls are not the problem. Troll droppings are: the replies.[...] the damage will be done before enough downvotes have been collected

My proposal was all about reducing the response to trolls by giving people an indication that they don't need to reply or vote. You speak of "the damage" as if it's an all-at-once situation. The troll threads of the last week continued to fester for days. What I propose would start a process of containment within the first hour (based on my observations).

It did occur to me to make certain posts un-reply-able, but that would be too problematic. My proposal does not impinge on freedom (except, in my alternate proposal, the freedom to continue to downvote a trash post, and that's one reason it's not my primary proposal). The strongest criticism of my proposal is that it's not needed or wouldn't work. Nobody has given any reason to think it would do any kind of harm.

---

Caution: Contents may have been coded under pressure.

Nice idea!

It tells you how you voted on it, and (if you've voted) displays its rep, but I don't see any other advance indicator. Double-check that and let me know what feature you're looking at.

---

Caution: Contents may have been coded under pressure.