What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 130 Stack Exchange communities.

The reward for answering difficult questions is disproportionately low. For example, I just got 90 points for a throw away answer that I googled for, but a total of 25 points for these four well-thought-out answers which required a degree of expert knowledge.

In some cases, the questions are old and the users are gone. In some cases, a "less good" answer has been chosen.

If there are no badges or points for this, where is the incentive? The feel-good factor wears old when you realize no-one is listening. What can we do about this?

Related-ish

share|improve this question
    
    
@Rich Seller: That relies on the answer being accepted I think? This isn't always the case. Besides, its easy to get a throwaway answer accepted if you're the faster gun in the west. Giving good answers is still an expensive way of getting those badges. –  Paul Biggar Sep 11 '09 at 22:45
    
This is why we really need badges around the "accepted answer" functionality. –  womp Sep 12 '09 at 6:18
    
The problem is even worse on SuperUser. Answer a question with "Have you tried VLC/7-Zip" and you'll instantly make 100 rep points. –  alex Sep 14 '09 at 14:25

5 Answers 5

Questions that require a higher degree of expert knowledge will inherently draw lower rep typically because the audience who can feel qualified voting on your answer is very small.

Anybody can upvote an answer on a simple question, that is why they get the highest audience.

I don't know how you would want to make an incentive to answering harder questions. We already have bounties, which will allow questions with lower views to attract more attention. This is slightly tangential to what you are looking for though.

Your problem here is that the system can't tell a hard question from a bad question. Both would normally not get the attention afforded to a better question. The only way this could be done is if there was some sort of extra feature, like a community rating on a question where you could go in and assign a "degree of difficulty" to the question. This degree of difficulty could then perhaps be used as a multiplier on the rep awarded.

I doubt this would ever get incorporated because I can already imagine there being huge gaming potential for it.

share|improve this answer
    
Questions need to be equalized so that there is always something worthwhile about answering them. If that can't be done with votes (which inherently relies on simplicity and traffic), then it should be done with badges. Every question has a possibility of an accepted answer, so that would be the perfect place to start. –  womp Sep 12 '09 at 6:33

I completely agree, and I do think this is a real problem. Three of my last four answers on SO (algorithm/graph tag) took hours of work, required detailed knowledge, are probably correct, and have gained me exactly 0 reputation. On the latest one, I don't even expect an upvote from the asker, since he doesn't actually have enough rep to vote! On the other hand, the 30s Google search for an obvious webpage was worth 65 points.

There are several problems. It's clear that these are low traffic questions. But it's also obvious that the few people that do view the question and answers don't vote because they don't feel they understand the details well enough.

For better or worse, I'm almost ready to stop answering difficult questions. It is genuinely rewarding to be able to contribute something helpful back to the community. But when most people simply don't know or care, it's hard to justify the effort. I do agree it is frustrating when there are one line regex questions out there which require zero effort and are worth hundreds of points (because they are "neat" and are easy to verify).

Other than specialised badges, I don't have any good solution. Perhaps all it reflects is that Stack Overflow simply can't be the place for all programming questions, particularly those that are more computer science. Specialised discussion lists (or academic papers) may simply be a more appropriate (if rather less fun) solution.

share|improve this answer
    
Sure stackoverflow can be a place for all programming questions. Just because you aren't getting rewarded with rep in the way that you would like doesn't mean that SO isn't the venue for it. –  TheTXI Sep 11 '09 at 22:04
    
You build a game with a reward system into a user experience, the user experience revolves around a game with a reward system. It's a fairly well-recognized principle in video games, but it seems to apply much more broadly. SO is designed to fail precisely in those areas where some folks would most like it to succeed. –  Mike Burton Sep 11 '09 at 22:08

OK, how about a "wall of fame". It wouldn't be for popular answers, but for really great answers. They could be nominated by anybody, not just the question asker. When they appear on the wall of fame, other people impressed by the depth and wisdom shown would presumably upvote them.

I don't have a solution for how to manage the size of the wall, since there would probably be tens of questions added per day, but hopefully not more. Perhaps it would be well to limit the size of it, or to limit the nominators to people with some points threshold.

It would also be useful if abandoned questions had a way of being taken over in some way, to award an answer. Perhaps people could "apply" to take over a question, and if it goes unanswered for a length of time, the take-over would be successful.

share|improve this answer
1  
I don't see how this would beat the bike shed effect. The nominations and votes will still come from the same pool of people, and they are still more able to comprehend not-too-hard to middling problems and solutions that the deep one. –  dmckee Sep 11 '09 at 23:10
    
@dmckee: it would come from clearly explaining the point. "this is for wonderful answers which are well researched, contribute expert knowledge, and havent gotten the credit they deserve" (or similar). –  Paul Biggar Sep 12 '09 at 0:14
    
@dmckee: perhaps we should restrict it to people with 1000 rep or so. –  Paul Biggar Sep 12 '09 at 0:22
1  
The bike shed effect gets everyone. Even really smart, broadly educated, highly knowledgeable folks can only be experts in so many areas. Even in as restricted a field of "Programming" no one is an expert in everything: Jon Skeet does not answer LaTeX questions. –  dmckee Sep 12 '09 at 1:46

I think tipping could help this.

If authors of the difficult question are particularly grateful for an answer, they could 'tip' the author of the answer with some reputation. Kind of like a retroactive bounty.

The tipping could be limited so that only the question's author can tip the accepted answer so that tipping abuse could be limited.

share|improve this answer
2  
Tipping would be another easily gamed item. –  TheTXI Sep 11 '09 at 20:44
1  
@TheTXI: is there a great difference in gaming a bounty vs this idea? If it were visible, people could report gamed tipping I suppose. What happens for bounties where people are cheating? –  Paul Biggar Sep 11 '09 at 22:48

Currently, whenever authors answer questions, their reputation is shown as a simple lump sum. Maybe it would be interesting to also display the reputation they have gained per tag used in the question. Something like this:

Eric [200], c# [100], winforms [30], ...

This would give users more detailed insight into people's reputation.

share|improve this answer
    
Wow. That would be some game changer. I think the villagers would be fetching the pitchforks for that one. –  Paul Biggar Sep 14 '09 at 18:57

You must log in to answer this question.

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