I’ve observed four kinds of answerers on Stack Overflow:
The Competitor
The competitor will defeat all the other answers on the field of combat.
The Helper
The helper wants to help the asker with their problem.
The Canonicizer
The canonicizer wants to make sure no one will ever have to ask or answer a related question again.
The Challenge Seeker
The challenge seeker is looking for interesting problems to solve.
It’s important to note each have notions of question quality. The competitor just wants enough to scratch out an answer before anyone else gets to it to eek out some sweet FGITW points. It can cleaned up to actually answer the question later.
The helper needs enough information to solve the asker’s problem, and is willing to work with the asker to solve it.
The canonicizer wants enough information to ensure that anyone else who has a similar problem will be able to find and recognize it, and be able to use the answer to solve their problem.
The challenge seeker just wants the asker to go away because the problem can obviously be solved in 20 characters of Perl.
All four of these can agree that “What is the correct Megaspam brand™ NATURAL ENHANCEMENT pill dosage for you? (Click here to buy.)” is a terrible question, but the SO process has pain points for all of them for any question better than that. How can we reduce that pain and provide the best experience for all our users working constructively together?
The Competitor: this is the type we want to discourage - defeating the other answers is probably not a good thing in most circumstances, and it creates an unfriendly environment. Albeit, there are cases where incorrect answers need to be challenged.
The Helper: Yes, yes, yes. These are the people we want to keep because they are friendly and extremely helpful.
The Canonicizer: Obviously these people are important and would keep our site tidy and organized, but we need to encourage our canonicizers to make sure they contribute in other ways as well.
The Challenge Seeker: This one is tricky. If the user actually gives helpful answers, then I say they can go ahead, but if they are just rep-whoring (yes, I’ll use that term), then we shouldn’t allow that.
As for working together; that’s tough. These are all obviously different people that have different ways of thinking, and they are going to clash sometimes. I think the best way we can mediate these clashes is by having a team of neutral moderators who not only encourage rule-following but also kindness and cooperativeness.
I didn’t understand the Challenge Seeker as a rep whore (that would rather fit the Competitor), but as someone who looks for challenging questions to solve, and who would probably try to solve them even if there were no way to post the answer, but certainly prefers to show the found solution. In other words, the Challenge Seeker would post even if there were no reputation, as long as the question is just the right amount of challenging.
In that interpretation, the Challenge Seeker would dislike the typical homework question, because it doesn’t pose a real challenge (as opposed to the rep whore who likes the easy rep earned by answering homework questions, and the Helper who may consider homework questions a great chance to help).
I think I understand the Helper and Canonizer, but not sure what you mean with Competitor and Challenge Seeker.
How does a Competitor “defeat” other answers? Do you mean he goes around downvoting competing answers? If so, that’s bad. However, if he does it by writing a great answer so that everyone upvotes it above the others, that’s a Good Thing.
Are you saying the Challenge Seeker passes over simple questions, but then really digs into interesting ones? OK.
Either way, a good voting mechanism should take care of all these types of answers. It’s the quality of the answer that needs to be judged, not the motivation of the answerer.
Or, are you asking how to keep each of these answerers with different motivations engaged? I’m not sure what your overall question/point is.
The Competitor won’t cheat, but will do whatever is allowed to get as much recognition as possible.
The Challenger Seeker will get frustrated if all the questions she can find are too easy. Otherwise you pretty much have her down.
The company I work for has millions of customers, so we try to cluster them into groups based on personality characteristics. Then we can talk about how features impact them. A big part of the question is whether these are representative groups. The other question is how to resolve the different definitions these users have for “question quality”.
Not sure if I agree with these. “The competitor” doesn’t really exist as a user stereotype, but is rather made up by “the poser”, who just wants to build massive rep and care about their public image, and “the tech nerd” who just wants to provide the best (or most complicated) technical solution. Both of these will post answers even when plenty of answers are present, “the poser” because they see a high traffic post and “the tech nerd” because they didn’t think the present answers were good or completely correct.
And then there’s plenty of answerers combining these two, suggesting a technical solution based on some obscure language feature just to show off their knowledge, not because it’s the best solution. Just look at the SO C++ community, it’s filled to the brim with people proposing obscure meta programming techniques using the latest (often mildly useful) language features as the solution to pretty much any question asked, just for the sake of using obscure features. Rather than to go with the KISS principle which is what a professional would do.
And then of course most answerers don’t belong to a single category, but are a mix of several.
This reminds me soo much of the flame warriors guide (and several other versions of forum characters/personalities).
Personally I am a challenge seeker and a helper.
The challenges I look for in the questions with bounty, and also I sometimes look for questions that are asked by high-rep users.
Helping makes me go against ‘the rules’.
Particularly when people ask unclear questions, then I still tend to figure out and try to help them. (with the result that my comments inquiring about the question might provide the answer and the questioner, often a one-time visitor, walks away with the closed question remaining sitting there not being improved and sometimes neither being deleted).
It’s nowhere that simple, of course. I somewhat resent that you imply the only explanation for the competitor category are two somewhat negative stereotypes.
I answered a lot of question. Yes, I wanted to be seen as an expert. I also like teaching, and think I can often explain things better than others. I often wrote answers even when there were other good enough answers because I thought I could do it better. Eventually over time, many of those answers got upvoted and rose to the top.
I did build massive rep, but that wasn’t the motivation. The respect of the community and being seen as an expert were. Rep was just one mechanism within the system to help with that. I never did something just because I thought it would gain rep, but I did like seeing the high rep score, particularly relative to a few other experts I respected.
In the end, the takeaway is that you have to give experts some recognition to keep them engaged. They aren’t going to judge individual actions by how many fake points each one will yield, but they do want to have some kind of “score” that one way or another shows them as being better than everyone else.
SO is an anomaly. We do have to consider how sites work when they get large, but the real target should be “normal” sites that aren’t so massive and therefore don’t have the skewed characteristics of SO. One solution is to break up large sites into multiple categories once they get to the unwieldly large stage. From my distant perspective, that should have happened to SO long ago.
Yep, simple generalizations are usually not that helpful.
It is true that rep stop mattering for many users when they go beyond the 20k or whatever you need to unlock all privileges. I stopped caring about it completely, many years ago.
I think maybe a system where you gain reputation in a specific tag or domain rather than generic rep could be a motivator. Like for example in electronics, if you got individual reputation/status in analog, power, RF, microcontrollers, PCB CAD etc etc I think that could be a better motivator? If someone wants to get “analog guru” status, they would have to answer a lot of questions in that specific sub-domain. It might even motivate people to learn more of areas they aren’t familiar with. Sort of like bronze, silver and gold tags work on SE but something more sophisticated.
I would consider the top 10/20 sites on StackExchange as showing the same ‘disease’ as StackOverflow and soon the symptoms will increase.
Many of the smaller sites also have:
(I have added images based on data from the electronics site as examples)
the ‘fastest gun in the west’ problem
decreasing number of votes, and decreasing scores on posts
discrepancy in votes distribution based on age
Less attention to older posts
To get back on topic. That last graphic particulary shows the strong shift to the ‘helper’ and the ‘competitor’. At least, in terms of voting there is only interest in the answers that are placed on new questions. The old questions do not turn up anymore in the voting carousel (that is the front page). You can see this also in the voting activity on particular questions. Often this happens in short bursts which indicates that this is related to the question being bumped to the home page (indicating that SE/SO runs on frontpage-activity and not so much on database-activity).
SE had badges for that, which never meant anything to me until I discovered that a gold subject-specific badge lets you close questions as duplicates single-handedly. Frankly, I always thought badges were kind of silly. What do they actually do for me?
Rep was useful because it was a light-weight attribute which therefore could be displayed liberally. Badges were klunky. That means it wasn’t as obvious who had what badges, which means there was no value in them. There were also too many of them to wrap your mind around, and some of them seemed to be contrived just so that they could be awarded as consolation prizes. I usually just chuckled to myself when the system awarded me another badge.
I’m not seeing that. It looks like the inevitable result of a large repository being built. The more stuff in the dusty archives, the less any particular piece of it will get noticed. In other words, some of these trends are just following “daily activity as fraction of total database”.
That should be no surprise. What this basically says is that the current activity is about the current activity. The vast majority of answers are going to be on new questions.
Good old answers do get regular upvotes. I think there are two types of question activity. Those that search and find something existing that answers their question, and those that don’t and write a new question.
Most people just blurt out their question with little research, but old questions do get found and apparently appreciated. I just checked, and I have the 18th highest rep in 2019 on EE.SE, even though I haven’t even logged in that entire time.
However, I don’t know what any of this means to the overall question at hand, which is how to accommodate answerers.
It is not that the activity of voting on older posts is just being spread out on more posts. It is the activity for all posts together which has gone down to nearly zero.
Those curves are for the total number of votes. And not votes per post (both questions and answers).
We have a growing number/bulk of posts, but after a week (or even a day) nobody looks at it anymore (and those few old top posts are not making much up for it).
That is why I say that the activity on the website is dominated by ‘frontpage-activity’:
helpers that answer very specific helpdesk type and homework type questions (which some see as detrimental because it encourages lower quality questions)
and competitors that search for new/hot/active questions in order to score.
(I have done this as well; while questioning myself for low scoring answers and searching specifically for popular questions that give potentionally high scoring answers; and yes it works, you can game the system in that way and ‘score better’)
The ‘canonicizer’, ‘challenge seeker’ and (I am adding) ‘quality controller’ are getting less ground.
It did not use to be like that. There used to be activity that was about older posts, past activity, as well (at least the voting). See also the following graph which shows the distribution when questions (on average) acquire their score as a function of their age (how much do posts increase their score in the first month, in the second month, etc).
(above is a graph for StackOverflow; see here for a graph of electronics.se)
See also On the value of scores and votes - #4 by MartijnWeterings for typical growth curves of the score of posts (you see that those answers acquire a lot of their score in the first day; in the old days questions still kept on growing their score with a reasonable pace, or even get into a polynomial growth; but nowadays the answers and questions on average get very little extra score after the first day and the pace by which they improve their score is decreasing)
Here we see indeed a dilution effect. In the early years of SE/SO a question would on average score more and more as the age of the post increase (except an initial drop after a short high peak in the beginning). But now questions only get their score in the first month (you can zoom in further and it is actually almost all in the first day or even first hours).
Questions (and also answers) do not last as long anymore. They only get one day or less to grow up before they are thrown into the big pool of old questions. It is becoming a rarity when they get a (up)vote or two in the process.
(It is also a difference in the style of questions. It is much more difficult to make a question like ‘the old days’ which has more canonical type posts. But I believe that the scaling and aging of the site plays a strong role as well.)
