I know that once a user passes 1000 reputation (or 750 in a smaller site), they gain the ability to see vote-counts, rather than vote total (upvotes - downvotes). However, I don't really understand what purpose this serves. I know that some activities are not given to a new user to prevent spam/abuse (for example, limited voting, commenting, etc.), but what's the harm in letting a new user see vote-count?

And also, it seems to me like certain abilities or more important than vote-counts, but received earlier. For example, a user with 500 rep is trustworthy enough to downvote, award bounties, access the review queues and some other things, but not yet trustworthy enough to see how many down vs up votes a question has.

Why is this privilege designed the way it is? Also keep in mind, I'm not trying to change it, I just want to understand the thought-process behind it, since it doesn't make much sense to me.


Why is this privilege designed the way it is?

Mostly due to it being a very data-intensive query. The score (upvotes - downvotes) is stored with the post. Getting the breakdown requires picking the data from the voting table (which is very big), and is therefore quite expensive (server resource wise).

Putting the ability to get this breakdown behind a privilege mitigates this by only allowing a fairly small subset of users to get this information.

| improve this answer | |
  • See also: meta.stackexchange.com/a/1007/262755 – Fish Below the Ice Nov 13 '15 at 18:56
  • 1
    If you are really curious you can always head to SEDE, if the question is from before last sunday... – rene Nov 13 '15 at 19:53
  • 2
    I thought the database model was fixed already to store each value separately? It's just four bytes per question... – John Dvorak Nov 13 '15 at 19:54
  • @JanDvorak - 10 million questions. And ~25 million answers. Not including deleted posts. It adds up. – Oded Nov 13 '15 at 20:06
  • 4
    35M questions * 4 bytes = 140 MB. I can't see that being significant in terms of storage space. I would guess it's less than one percent of the storage used for answers, not counting comments and deleted comments. I can see there being an issue with trying to compute all of these at once, but perhaps computing the split totals lazily over time could help? You have to compute the splits anyways (for some transient time the splits could be stored but not available to the commoners), so why not store them? – John Dvorak Nov 13 '15 at 20:17

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