5

According to this question, reputation is a "broad aggregate representation of the trust the community puts in you."

However, it seems to me that the more "popular" a platform is, the more likely your answer will get seen, and the more likely it is to be voted on.

Your "reputation" then, is a function of both the trust the community places in you, and the size of the community that is interested in the same thing.

It is entirely possible for someone with deep knowledge in domains that most programmers care less about to have significantly less reputation than someone who has modest knowledge of a more popular subject. Note especially that the questions that generate the most reputation seem to be generic opinion based queries. The kind that often generate flame wars because people are passionate about them.

I am not saying that such questions are necessarily bad, my point is just that reputation has a strong dependency on the number of people interested in looking at the question, and a much more tenuous connection to the actual competency of the answerer.

8 Answers 8

3

It's subjective. It isn't too subjective.

Of course, what it really is is XP.

Let me recapitulate here something I said in a comment to someone who was being all snide about how if you care about rep you're a bad SO user:

Dude, you don't get it. SO is a webgame, like all these generic productions where you click on things repetitively to level up and whatever, except you accumulate XP -- oh, I mean rep -- by helping people solve problems and contributing to the sum of public technical knowledge in the world. Taking the hoarding impulse, surrogate power drive, and general OCD that make people play computer RPGs and harnessing it for the forces of good, and moreover productivity, is SO's great innovation and victory.

Calling it reputation is just a smart bit of social engineering.

3
  • 2
    I think this is probably the best answer, because it's closest to reality. SO is a webgame, and reputation is XP. For reputation to be more trustworthy and useful, it would have to be weighted based on tag to some degree. Right now, it's just trying to level up. Aug 6, 2009 at 2:54
  • What does the abbreviation "XP" mean? (in this context) Sep 2, 2009 at 9:11
  • Experience points.
    – chaos
    Sep 3, 2009 at 1:53
7

Even though reputation is biased towards certain questions, it works out in the end. Users tend to be active under specific categories, and as such reputation serves as a good relative indicator of experience. Everyone who posts in say, Java, has the same relative advantage and thus reputation can still be used to rank users under a particular tag.

I also must stress that reputation isn't the be all end all of a user's worth, there are tons of very intelligent people with low reputation. It's a relative indicator, and if you take it as "this user is better than that user because they have more rep", then you're over thinking it.

I close with a somewhat general agreement, it's good enough.

2
  • 4
    Exactly. My rep on SO is higher than my brother's, and he's been coding much longer than I have, and is orders better than me. He just doesn't spend as much time hanging around taking in the rep. Aug 5, 2009 at 2:56
  • I disagree that things even out. Mostly because a person may primarily comment in one area, but they may also make forays into other places. I think that because reputation displayed does not actually represent their success in the category the are answering it is too subjective. Perhaps displaying total reputation is fine, but next to that a weighting based on how the rep. breaks out based on the tags for the question too. Aug 6, 2009 at 2:58
4

Overall, I consider reputation a reasonable metric, especially for those above 10000. People with this kind of reputation demonstrate persistence and a reasonable level of knowledge and skill. However, reputation is not without its problems. All other things being equal:

  1. Answers for popular questions (having a higher number of views) tend to get more rep.
  2. Funny answers or answers that resonate with a particular "group" (test-driven developers, for example)tend to get more rep.
  3. First answers tend to get more rep.
  4. Answers to difficult questions, which should be rewarded with more rep, tend to get less rep because it is difficult for others to determine if the answer is really correct.
  5. Rep is highly sensitive to how much detail is given in the answer. Too much or too little detail, and rep decreases.
  6. Answers to clearly-worded questions tend to get more rep.
  7. Answers to questions that have no clearly correct answer will get more rep, as those who agree with their own opinion vote up the matching answers.
  8. Answers to controversial questions will get more rep.

I offer for your perusal the following question. This question took me literally ten seconds to answer, and perversely returned 280 rep.

C# function to find the delta of two numbers

This question continued to return rep for a relatively long period of time because the community didn't take the question seriously, and began adding joke answers (the question itself is at -3 votes), which pushed it to the top of the stack on the front page several times as an active question (the question itself has been viewed 465 times). Of course, I was all too happy to fan the flames by asking if anyone had the recursive answer, the linq version, etc.

Don't even get me started on the whole Community Wiki/Rep-Whoring thing.

1

Yes.

4
  • Aren't you glad you made this community wiki? Aug 5, 2009 at 2:08
  • 1
    The main reason it's getting downvotes is probably because it is a wiki. Aug 5, 2009 at 2:10
  • 1
    Just too tempting, eh? ;-)
    – Shog9
    Aug 5, 2009 at 2:24
  • I agree. Isn't it clear if there's no one to judge the answer - there will be no votes. It's just an extreme, in reality lesser topics are simply voted less frequently.
    – Evgeny
    Aug 5, 2009 at 4:04
1

My take on it is that is the intent. By its very nature reputation and the scores given questions and answer are a subjective measure by the community. There is really nothing objective about it. So, when you ask the question, is it too subjective? I think you miss the point that it is and is intended to be completely subjective. But, as in parimutuel wagering, the collective consciousness tends to do quite well.

1
  • Actually, recent research indicates that most people, even in large groups, do a very poor job at making good financial decisions. See the August Scientific American. Also, see the recent world credit crisis. Aug 6, 2009 at 2:48
1

Reputation is a value that roughly reflects how many people trust or value the person. That's what the "broad ... community" really translates to: population count => popularity.

1

The question people should be asking is "too subjective for what?".

Things you can know about a person with high rep:

  • Knows at least a bit about the subjects he gets more points on.
  • Dedicates lots of time to the site. (So he cares about the site and is likely to make a good moderator.)

Things you cannot know about a person from his rep:

  • His intelligence.
  • His suitability for a particular work.
  • His level of expertise in any particular subject, at least for regular cases.
  • His communication skills (for that you have to read his answers.)
  • His hair color.
  • Gazillion other things.

So mostly the answer is yes. But the answer is also no if you care about the things you can deduce from the rep.

2
  • Yes, but rep isn't topic-specific. Some guy who knows tons about .Net stuff might make some comment about Erlang and be wrong. However, because of his massive reputation, his answer might get more weight and/or respect than it deserves. Aug 6, 2009 at 2:50
  • But you can tell where most of his rep comes from.
    – Vinko Vrsalovic StaffMod
    Aug 6, 2009 at 22:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .