12

A question was posted to SO today, which I answered and then received a number of upvotes on. The question was later migrated to Programmers.SE, where it continued to accumulate votes. The votes received on the new site contributed to my reputation, but those received from Stack Overflow did not.

  • is this intentional?
  • are we breaking expectations by having the actual reputation earned from a question be different than its visible number of upvotes (ignoring the existence of CW)?

(Naturally, the reputation earned on SO disappeared from my audit trail after the migration, and after a reputation recalc will disappear from my total score, so I'm not asking about this part. I'm only asking about the deficit of reputation on the new site.)

Improve this question

2 Answers 2

Reset to default
8

This used to be by design, but has now been amended. Below is a report of how it used to be.

As noted in this question about the reputation audit, it's been stated that the reputation is intentionally not migrated, as the votes on the post are not owned by users on the target site. Since no one on the target site has actually upvoted you to consider your post useful, you do not earn reputation.

It does somewhat break expectations, but migrated questions also have non-attributed votes in the first place, and it is possible to double-upvote someone using that.

Improve this answer
5
  • I can live with that answer, and I understand why it's necessary. But it does suck when an on-topic question is asked on SO and then migrated to programmers because it's a slightly better fit, and as a result you lose 500 reputation (even though your accounts are linked on both sites). I feel like we should be more careful and judicious about migrating questions with high-rep answers, particularly when the question is on-topic on both sites.
    – Mark Peters
    Feb 28, 2011 at 14:26
  • see my answer below (or hopefully above later, sorry!) :)
    – Nick Craver
    Apr 19, 2011 at 23:46
  • @Nick You don't have to apologize for invalidating formerly and solely mechanically correct information with better and popularly appreciated and desired functionality in the system. ♪ Really, I'm pretty sure this change is something everyone will like.
    – Grace Note StaffMod
    Apr 20, 2011 at 12:46
  • @GraceNote I just found this thread, and I don't particularly like it. It's minor though, the real problem is on the target site.
    – Gilles 'SO- stop being evil'
    Sep 12, 2012 at 15:42
  • @Gilles With over a year of retrospective on this and having seen the impact, especially on SE 2.0, I don't disagree with you on it. Votes alone already cause a terror in some of those scenarios.
    – Grace Note StaffMod
    Sep 12, 2012 at 15:48
9

This was not intentional, it was a bug. You should get all rep from the migration, from the source and destination sites. You will also keep the rep on the source site, unless the question is deleted down the road (the same as any deleted content).

You may have seen some "phantom rep" questions here on meta about users mysterious gaining pseudo-random amounts of rep spontaneously. Well...guess where that rep came from? You guessed it, a bug in migrations (or rather a Linq2SQL entity reference change) was causing the user to be pulled from the wrong database (the source rather than the destination).

The votes, etc were being inserted correctly but the intended user wasn't actually having their reputation increased as a result (though a recalc would fix this). We issued a recalc last week of affected users, but the root issue was only fixed last night...so we'll be recalcing the remainder of the affected users tonight.

Improve this answer
19
  • @Nick, just to get this straight: If a question with, say, 5 upvotes is migrated, then the user keeps the 25 rep on the source site (assuming the question is not deleted) and gets 25 rep on the destination site?
    – Hendrik Vogt
    Apr 20, 2011 at 16:17
  • @Hendrik - correct, since they have 5 votes on each site. Keep in mind that closed (migrated questions are closed) are eventually deleted, so when this happens it'll be as if the rep "moved" instead of was duplicated, that's ultimately a temporary situation.
    – Nick Craver
    Apr 20, 2011 at 16:27
  • @Nick: Thanks a lot for the clarification. I just realize that I have two more related questions: Does the same apply to answers? And is the rep cap on the destination site taken into account?
    – Hendrik Vogt
    Apr 20, 2011 at 16:30
  • @Hendrik - yes and yes, the votes are migrated as if the voter (or community, if the voter doesn't exist there) made them...all of the rules, caps, etc apply.
    – Nick Craver
    Apr 20, 2011 at 16:36
  • 1
    @Nick: Great, thanks. Now I have to add that I don't really like it: If, e.g., an answer with 50 upvotes gets migrated from SO to tex.sx, then the user will get 500 rep just like that. (Or are the 50 votes counted as cast at migration time?) This would be a lot harder on tex.sx! (Just want to mention this here; actually it would be stuff for a new post.)
    – Hendrik Vogt
    Apr 20, 2011 at 16:41
  • @Hendrik - the votes maintain their creation time when migrated...the rep you have on the initial site effectively transfers over, like everything else related to the question...
    – Nick Craver
    Apr 20, 2011 at 17:07
  • @Nick: Thanks once again for filling in the details! (Now I'm convinced that I don't like it ...)
    – Hendrik Vogt
    Apr 20, 2011 at 17:09
  • @Hendrik - Can you tell me what you think is not fair? You get all the rep from upvotes on that post, is there any confusion there?
    – Nick Craver
    Apr 20, 2011 at 17:11
  • @Nick: It's mostly what I wrote above: On SO it's not that hard to get 50 upvotes on an answer, whereas on tex.sx that would be exceptional. One more point: The post was upvoted by a different community with different standards, so maybe it was considered a good answer on the source site, but will be considered a mediocre one on the destination site!
    – Hendrik Vogt
    Apr 20, 2011 at 17:17
  • @Hendrik - If the question isn't closely related to that site, that's a different issue IMO...it shouldn't have been migrated in the first place.
    – Nick Craver
    Apr 20, 2011 at 17:23
  • 2
    @Nick: I do mean questions that are 100% on topic for the destination site. It's rather easy to give an example: Go to the tex.sx questions, sorted by vote. The top question is CW, so forget that one. The second one is a genuine excellent tex.sx question with 29 favorite markings. The third one got migrated from SO and wouldn't have attracted that much attention on tex.sx. It's just some (slightly above) average question (1 favorite) - but the top answer has 71 upvotes. Which for tex.sx is more than exceptional.
    – Hendrik Vogt
    Apr 20, 2011 at 17:31
  • @Hendrik - So it would be more fair that those answers got 0 votes, 0 rep for their effort? :)
    – Nick Craver
    Apr 20, 2011 at 17:39
  • 1
    @Nick: Good question. For that answer with 71 upvotes, I'd say yes, indeed. Already 20 upvotes for that answer would be a lot more than I'd ever expect; maybe around 10 would be normal. And 0 is a lot closer to 10 than 71!
    – Hendrik Vogt
    Apr 20, 2011 at 17:42
  • @Nick: And there's still the business with the different standards. But of course you've made a point, 0 rep doesn's seem fair either. (Let me point out that the old system was all votes, 0 rep.)
    – Hendrik Vogt
    Apr 20, 2011 at 17:44
  • @Hendrik - Correction: 0 rep until a recalc, a recalc would have (and did actually) fix these migrated situations. I think you're worrying about a rare case IMO.
    – Nick Craver
    Apr 20, 2011 at 17:47

You must log in to answer this question.

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