>> This feels deeply unfair in some sense... Having said that, I’m not sure what Stack Overflow should do about this.
SO would benefit greatly by implementing a (relatively slow) decay function on all its reputation points. Technology inevitably moves forward, a great answer (or question) from 6,8,10 years ago is the most likely age to have a huge quantity of upvotes, while also most likely to be obsolete.
While rep alone shouldn't be the primary motivation for answering questions, when people see their point totals heading south after a few years, it might provide an incentive to get back on and reverse the negative flow by answering new questions and/or updating prior answers.
I'd also like to see the ability to re-ask questions after a few years without them being dismissed as duplicates.
There have been a few questions where the answers no longer worked due to being obsolete, and I had to figure out the new solution myself. And in some cases, the answer never worked properly but was accepted as the best known at the time, and I had one that worked. But there was no point posting, because the accepted answer from X years ago dominated, and it was unlikely that a contribution from little me would be noticed.
In fact that's why I never ended up joining SO despite being a heavy user - on each of the occasions I felt I could make a useful contribution, it happened to be a question that was old and "settled", and I felt it wouldn't be worth the effort.
I never thought about that. I actually think this is a great idea.
One big issue is a lot of people would be upset at the idea of their reputation getting smaller.
Maybe have a new "influence" or "helpfulness" score along with reputation that takes into account the decay you are talking about. Or maybe just have a different scoring mechanism altogether based on how knowledgeable and helpful your are to the SO community.
The thing I wish for most would be some mandatory versioning on the question and answer. If you are looking for a solution to Android 4.4, answers that are specific to 6.0 are worthless. I know it won't be 100% accurate but if people put in their best guess it could really help with filtering. I know they can put it in the title but you get mistyped and different names for things and I'd rather the tool make it mandatory but also quick and easy to just pick with just a few clicks.
An answer will already decay naturally if it goes out of date as people will downvote it for being wrong. There are plenty of relevent 10yr old questions and answers. Decaying old answers will just make good info harder to find.
imo they should get rid of the gamification. change the upvote to a "thanks" button and don't so the totals
A decay function on reputation might help, but I think the real issue is treating reputation as a precise indicator of user quality. But is a user with 40k rep really meaningfully better than a user with 35k rep?
An idea that I had was to hide reputation entirely from users and move to a flair system based on fuzzy metrics[1]. If the signal isn't precise, find ways to reflect that when displaying it.
[1] I imagine this as a system independent from badges, which are permanent and triggered by certain milestones/events.
I am more of thee type that creates a new account for answers/question from time to time, because I think the rep system is ridiculous as it is and I simply forget my login details. Neither do I know any names on SO, nor do I even look at the rep someone has that answered a question. Not being able to comment to get specifications to questions is bothersome, but at least the barrier is low.
This happens anyway, because the number of points in circulation increases. 10 years ago, having 20K points on StackOverflow was both more rare and meaningful than today.
If an answer is no longer relevant, it will, stop being upvoted, possibly even downvoted. So it’s score will become lower relative to newer, more correct answers. At least that’s the theory. The problem is the discouragement of asking new questions, the pointlessness of a new answer on a 10 year old question with 1,000 upvotes, search engines sending people to out of date answers, and that the upvotes don’t account for knowledge and experience. I have no idea how to solve those problems.
I think you misunderstood GP. My take on their point is that when a question from 10 years ago about a C#/SQL issue has a top answer with 300 points about using a library that’s 12 years old, but a better answer with Linq from this year only has 4 points, people are probably going to either go with the 300 point legacy answer. A decaying function can implement relevancy.
No, that was perfectly clear. That's why I said people should not use reputation or upvotes to judge anything in the first place including how good an answer is. They should read all the answers (including the more recent ones) and if needed do further research to find out which one is more updated or relevant at present. I do it all the time particularly for questions on things like git and even programming languages like python which have lot of better way of doing things in last 2-3 years than say 10-15 years ago.
In addition to being offensive (which, mind you, goes against HN's guidelines), you're literally agreeing that the current way scoring works is bad for finding up-to-date answers (of course "relevance" is more vague). I'd say any solution that makes it easier to automatically push better (in this specific case, updated) answers to the top is a win, especially for such a widely used system.
> you're literally agreeing that the current way scoring works is bad for finding up-to-date answers
No you're "literally" misinterpreting what I said. I said there should not be any scoring system at all and people should select an answer based on its own merit, usefulness and how recent they are as per their own judgment instead of relying on other people. In fact this should be true for anywhere on internet including this site.
SO has a rep system. I offered an idea that I think would make that system more effective. On the other hand, if you think the concept is misguided, don't give a flying f* about it, and think it should be eliminated, you're welcome to suggest that to the people who run the site.
How would it change the UX (or UI) at all? Question and comment "scores" constantly change as they get upvoted (and occasionally downvoted). This would just be another factor in determining the present score.
As far as "theory" questions, such a designation could simply be another moderator privilege -- high-rep individuals would have the ability to vote to prevent a question from decaying. Perhaps an additional requirement would be that the question must have been transferred to Community Wiki.
You haven't explained why "this would be terrible UX", and it would be fairly trivial to add a special "theory" tag or something more general (maybe restricted to high-rep users or mods) to avoid the problem of unwanted decay.
> Because it seems obvious that watching one's hard-earned score shrink would be a bad experience.
But it's a Q&A site, not a point-gathering exercise. Leave the displayed score of the answers/questions but decay the benefit the author gains from it.
Indeed, but if you've put in a lot of work asking/answering good questions and they reward you with standing then it sucks to suddenly start losing that. There is already a certain level of inflation in that those who are continuously active will keep raising the bar. I am rather proud of my contributions and score on SO and don't want to see that go away
SO would benefit greatly by implementing a (relatively slow) decay function on all its reputation points. Technology inevitably moves forward, a great answer (or question) from 6,8,10 years ago is the most likely age to have a huge quantity of upvotes, while also most likely to be obsolete.
While rep alone shouldn't be the primary motivation for answering questions, when people see their point totals heading south after a few years, it might provide an incentive to get back on and reverse the negative flow by answering new questions and/or updating prior answers.