I used to have the title taped above my desk in my lab office, to remind me that not everything I think is interesting to others, and not everything that’s true matters in every discussion¹. Of course, interestingness is not the most appropriate criterion for a comment in every situation (the hurricane is coming, but people, what would be the most interesting thing to do?), but it’s what I look for most in online discussions at the moment.

To maximize the “interesting-ness” of conversations, I suggest that too things are beneficial: a common understanding of what each member knows and enjoys, and an influx of new things that are not known to most members, but can reasonably be believed to matter to them.

By a common understanding, I do not mean a large shared pool of knowledge, but it does help to know what’s been said before. This seems to require a small, close-knit group, or at least a group whos state of knowledge lies in a pretty tight distribution.

The requirement of a constant influx of new ideas would seem to contradict the first, but it actually depends on it: if I have no idea what’s already known, I won’t know whether I’m introducing something novel, or harping on an old refrain. This seems to require growth, at least in what each person learns, or in the addition of new members.

The push and pull of close-knit groups and the need for new members or ideas leads to a problem for web communities. Growth is probably necessary, but too much destroys the ability to have a meaningful representation of what members know. As a result, boring comments are made, over and over again. To a new member, these comments often provide value and insight, but to older members, the comments are repetitive at best, and karma-whoring at worst.

However, if some web communities will continue to grow, it might be useful to have some guidelines. As a heuristic, I’ve developed a system I try to keep in mind when making comments, to rank or categorize them². I tend to think of these comments as if I were making them at a scientific conference, so I’ll place them in that context. Sometimes taking what happens in an internet community and asking how it would look in person offers an emotionally salient perspective.

These are not rigid, but it might be a helpful starting point. Incidentally, I believe they’re most helpful in deciding whether to upvote someone’s comment, or otherwise reward it. There are other reasons to upvote, but in a community where the same ideas make the rounds over and over, I believe interestingness is often more appropriate than most other considerations it’s in tension with.

I0: One-line insults

These types of comments usually don’t address anything at all about another comment, a submission, or any idea in general.

“what a doofus”, “I am so fing tired of this crap”, “why would anyone think that?”

If I did this at a scientific conference, I would be met with blank stares, and be shut out of any good discussion.

I1: Comments on procedure or minor points of form.

Usually these points are either obviously true, or subjective. They can be interesting in the appropriate context, like when someone asks you to critique an essay. Otherwise, they’re just boring.

“I don’t like the font choice”, “why is this being flagged or downvoted?”, “it would be great if we could change the title”, “why won’t this work without javascript?”, “why do the numbers start at 0 instead of 1?”

The danger is that most people might agree and all have strong feelings. The font choice might really be hard to read, the title might be very misleading. So in an internet community where everyone can vote, or even in a social circle where people can nod their assent or give other nonverbal feedback, it totally derails any other points the article or speaker may have made.

You can usually spot these beginning with the statement “I hate to be off-topic, but …”, or “It’s 2013 guys, can we stop …”

At a conference, these might be relatively harmless, but if I have 5 minutes to speak to someone, what have I learned by telling them “gee, I like your poster, but let’s not talk about that; let’s talk about your appreciation of fonts”?

I2: Calls to action

Often these comments do reflect the beliefs and knowledge of the community, but are rarely novel. The most popular ones make a statement a substantial portion people in the world at large agree with, but often not all, so that those who agree feel a need to upvote, or nod, so that their perspective or point of view will gain more visibility, and ultimately, “win”. They often omit reasoning, but sometimes include it.

Sometimes these comments are useful, in times of crisis, or when decisive leadership is needed. Other times, they’re never going to convince people, so they serve only to make the people who already agree with you like you more.

“God is a delusion”, “Science funding should increase”, “We need to stop X”

Often I agree with these statements, but if I attended a conference where the highest honors or best invited lectures were given by people who said nothing but “Science is important, guys” I would stop going.

I3: Admonition

When a user does something mean, or uninteresting, sometimes another member will submit a link to community guidelines, point out the error. This is sometimes helpful when the admonisher is a respected member of the community that helps set community norms. Often times, it just serves to gratify the ego of the commenter, however, even when they’re right. In person, people who do mean things begin to be left out, and specifically taking them to task just legitimizes whatever they’ve done. This type of comment can be useful, but it is rarely interesting.

At a conference, if an older member gently, and often in private, points out that a particular behavior is not generally encouraged, it can do well to regulate the community. On the other hand, if the chair at a speaking event humiliated a commenter who was legitimately mean, it might be well-deserved, but it doesn’t really make the conference more interesting.

I4: Establishment of Fact

This kind of comment can be very interesting, but often is filled with unecessary details, as if the commenter feels compelled to detail every single thought they had about a submission or idea.

“1) x,y,z 2) I like p,d,q in the author’s paragraph 3, but I would change q slightly to q’ … N) Lastly, they completely forgot about d”, it doesn’t change their story much, but why not mention it in a 5 paragraph blog post?”

The most important way to make this comment interesting is to ask if a) your fact is unlikely to be known to the audience and b) relevant to the submission.

I5: Association and Expansion

Sometimes when you’re confronted with an interesting point that requires technical knowledge, it’s tempting to expand from talking about that small, hard thing itself, and discuss only its broader implications, or something else it makes you think of, that “hey! guess what! I know little about X, but lots about X’ let me talk about that”.

“Expanding the comment” with context is not always bad. The most interesting points are placed in some sort of context, but only as a general reminder. This is one place where close-knit groups can help. Everyone working on quantum computation knows how important it is, so they don’t need to spend a conference discussing where it lies in the grand scheme of things. The grand scheme is very interesting at first, but once it’s established, it’s not so much fun to repetitively bring up, unless some new development inside the field (or outside!) shakes it up considerably.

Likewise, associations to other ideas can be extremely interesting, but only when they don’t ignore the original point. An example of the latter might be:

“Whoa, it’s great that you might be getting nuclear fusion to approach net-energy > 0. That reminds me of when Fermi first established the self-sustaining nuclear fission reaction. He did it underneath the stadium at the University of Chicago’s football field, which had formerly been a football powerhouse. It really emphasizes a transition from sporting to academic excellence …”

It’s not a bad comment, but it’s not as interesting, perhaps, as an association that points to a potential change in the topic at hand, or how the topic at hand will point to a change in some other field: “whoa, that’s cool, combined with this other group’s results there could be a working nuclear fusion reactor powering SF in the next 10 years. Here are some obstacles they’d have to overcome, but here are some ways they might do it!” It definitely branches and explands, and it even broadens the topic, but it does so in a very interesting manner, provided the above fact is not already well known to the rest of the audience.

I6: How the world will change

The most interesting comments illustrate how the major perspective put forth by a submission or other comment will alter the way the world is organized, or how we understand it to be. Or they change the way I think or approach the world.