I noticed this within the context of submissions and /new when I posted two blog posts of mine. Both posts were x000 words so minimum reading was over 3-5 minutes (average time on site was 1.5 minutes so not everyone read to the end).

Both times neither posts got any upvotes until half way down the first page. They were surrounded by topical submissions, news or etc. Stuff that either had a simple message in the title or was only a few paragraphs long.

Except a strange phenomenon occured when the posts fell to the second page, readership doubled. Whereas on the first page Google analytics real time reported 10-14 active readers this jumped to the low 20s. By this time the posts had received 5 upvotes. On page 1 they were maybe the 4-5th most upvoted submission but once they reached the 2nd page they stole 1st or 2nd place among their non-frontpaged peers.

The lesson I took away from this is that short or topical submissions have a strong advantage. It is easier and quicker for someone to decide to upvote. In same cases they may upvote without reading at all if the title is descriptive enough. Meanwhile people do like longer posts but reading a full essay is barrier to upvoting.

I don't suppose this is already being done, is it?

For example, there is an increasing tendency to create dummy accounts for the sole purpose of dissing another user without getting your own account dirty. I was recently at the receiving end of this, and I have since noticed this happening more often all over HN. The solution for this would indeed be simple: don't allow new users to post comments for, say, 8 hours. In general these kinds of measures aim to make destructive behavior less fun, and I think they do work to an extend.

Then there are problems of style that can be solved by having guidelines.

Getting rid of "lolwut", "ORLY", or "Pro Tip:" comments is something we should all agree on; yet those are regularly upvoted, as are remarks that personally attack other users. This can be solved by declaring a ban on these kinds of things, because I believe the majority of users would honor those guidelines of they were made clearer.

Of course, the larger problem is that this even needs to be stated officially in the first place. It should be obvious that lolspeak, at hominem attacks, ridiculous trolling, and employing the basest rhetorical fallacies is not something that we reward here on HN, but nevertheless voters do reward them. That's indeed a problem of audience, not something that can be completely addressed by programming or guidelines.

Then there are people who see destructive criticism as the main problem, which I don't agree with. Personally, I try to keep my criticism constructive if possible and I do appreciate others doing the same. But when I want to receive feedback, I'm glad about any kind of response really as long as the tone of the discussion remains civilized.

The worst thing that can happen with feedback is not getting destructive criticism, it's the absence of any reaction at all.

But I agree with you. You don't fix human problems by adding technical controls. You fix human problems by adding technical measures designed to enable human controls.

My proposal is relatively simple: figure out a high karma number (high enough few people have it) and allow these people to issue an incivility warning. If you get more than a certain number of incivility warnings in a month you get banned, first temporarily and on the second time perhaps permanently. These reprimands would be public.

People need feedback. People need to know when they are stepping out. One of the real concerns I have with hellbanning (though I recognize in some corner cases it may be useful) is that it denies people this feedback and therefore does not encourage people to stay within the lines of acceptable conduct. So make hellbanning only for cases of people who are long-term repetitive problems or spammers. Implement a system that gives feedback to users for the rest. And allow the most senior people to enforce civility constraints.

Technology needs to empower people to manage people. Software should not manage people.

I agree, in a way.

The only communities I have seen that associates numbers with human merit or trustworthiness are the ones populated largely by programmers. I feel this is telling.

I always have low scores on these sites because I am just not motivated to contribute to them. While Jeff Atwood was on to something when he said that users can do a good job when they feel invested in the quality of content on a site, we just cannot expect numbers representing reputation or consensus to always be set rationally. That, and I just hate thinking I will be assigned a score for trying to contribute to a conversation when I have no idea how the site's culture will interpret what I do. We may as well carry those 1-10 judge cards around our necks and stick gold stars on outsiders' foreheads. It's absurd... but apparently it's necessary. It shouldn't be, but I know I can't handle 1 million+ users as a moderator.

One of my old jobs was to facilitate a chapter of Socrates Cafe, a national discourse organization. We allowed anyone at all to come in and share his or her experiences. My job was to facilitate the discussion and put out flames fanned by conversations between opposing parties. So, when atheist/theists, right-wing/left-wing wars erupt, I was supposed to take varying measures to restore order without evoking enough negativity to scare away members. That is, I had to try and settle everyone down while making them feel BETTER about everyone else.

I cannot begin to tell you how ridiculously hard this can be, and technology does not help unless you use it for shameless enforcement.

Intuition and context-awareness helps tremendously, and there have been times where I simply could not think of any solution to one particular heated debate outside of, no joke, an air horn. I hated having to blast all of the words right out of the room, but the group we had (to me) represented a community gone wrong that needed flat out intervention. We never once used tools to try and change the rules of conduct. Instead, we used technology as a brutal and unapologetic last-ditch enforcement of existing rules.

Now... That's a group where everyone saw each other in person. Clearly not the same thing as an online group. That's worse. In this setting, technology is the only thing you HAVE to mediate all interactions, even the "human" ones.

I don't have a real solution to the sorry state of large, rude online communities. In fact, if you made me a moderator, I would probably go on a banning spree until everyone was too terrified to stay on their own accord.

Even so, but I doubt anyone else really has a solution either... Including the OP. Improvement can only be measured until online cultures change in a way to push newly set boundaries.

Recently I have started to realize just how wrong this is, first in reading what others have written on IEEE spectrum on the so-called "Automation Paradox" and later internalizing that business systems usually treat people as integral components. Even if humans are relegated to a supervisory role it is important to given them enough to do to maintain the system and failure to do so in vehicles can result in air travel catastrophes or ships running aground. But the same holds true for every other field. There is such a thing as too much automation. Especially where we are dealing with social environments, like HN, that human touch is what is most important.

I think in this the first question is what should humans be doing, and the second is how can the computer system enable them to do it. Unfortunately engineers tend to forget the first question and so never reach the second.

(In contrast, pg's idea to assign different weights to comment votes would be invisible, leaving the site UI as is. That is a purely technical solution.)

I didn't mean to imply that technical solutions wouldn't make a difference, even dramatic ones, just that they could not completely take the place of some manual effort.

Because by "user interface", we mean "human interface" or "people interface". Such changes are human-/people-oriented, pretty much by definition.

pg's "solution" is purely technical, since the human interface wouldn't change at all.

1) Don't you think Wikipedia administrators and Stackoverflow mods fulfil that "gardener" role? (I think so).

2) Consider that above "gardeners" are often criticized harshly on HN (even though, in my opinion, they are a positive force on Wikipedia and Stackoverflow). Don't you think strong moderators would have a bad time on HN?

    >they are attempting to fix a people problem using technical solutions

Don't get me wrong. I'm glad we're not throwing "technology" at the problem. But I'm glad for a selfish reason. I think adding features and modernizing the site would draw attention and users, probably the "wrong" kind.

Yes, and I cringe every time I see the next startup come in trying to do just that. The reality is the people putting money behind this whole movement don't see it as you and I may see it. So as long as people are making money and they continue to prey on people's cognitive biases then we will continue to see this happen.

I just picked this particular thread because it was new. The other one was too busy and already halfway down the front page; the same comment there would have worse chances of actually being read.

If all you have is a hammer, everything looks like a nail.

Don't have anything else constructive to say.

Of exilelifestyle.com fame?

Let us not forget the object lesson from Digg's big redesign.