8

Yesterday, a fairly new user asked a question in one of the tags I follow: read excel file(which is in classpath ) via apache poi

About 18 hours later, they asked pretty much the same question again: read excel (which is in classpath ) via apache poi

Virtually the same title (modulo one word), almost all the text the same, code the same etc.

I flagged it as a duplicate, and a mod later came through and zapped it, but it was a non-zero amount of effort for the community

Could we not have the system automatically detect these kinds of very very similar duplicate questions from new/new-ish users, and tell them to edit their question rather than re-asking it?

9
  • 6
    One word : how? – John Dvorak Dec 11 '14 at 23:00
  • This change would be a huge problem on Mathematics, given that users aren't too creative in naming their questions. I can't tell you how many times I've seen "Integration question" or "Probability" as a question title. – HDE 226868 Dec 11 '14 at 23:44
  • 3
    @HDE226868 Identical question titles are already blocked. This is why one sometimes sees "Probability question 3" and so on... – user259867 Dec 11 '14 at 23:50
  • @Behaviour Thanks. Hmm, I suppose I'm misremembering, then. – HDE 226868 Dec 12 '14 at 0:04
  • 1
    @HDE226868, why would it be a problem if this change happened to (as a side effect) block questions with a poorly chosen super-generic title? That sounds like a feature, not a bug, to me, especially if it provides feedback that prompts the asker to provide a better title. That will make the site more useful to others and increase the chances that the question gets attention from folks who might be able to answer it. If that was the side effect, it sounds like it'd be a win-win all around! – D.W. Dec 12 '14 at 1:58
  • @D.W. I wasn't proposing blocking those titles - I thought it would be a bad thing to block them! – HDE 226868 Dec 12 '14 at 1:59
  • 1
    @HDE226868, I know. I think you've misunderstood me. I'm asking why you think it would be a bad thing to block super-generic titles like "Probability". Maybe read my comment again? – D.W. Dec 12 '14 at 2:00
  • @D.W. Definitely misread, thank you, though I think your revisions made it clearer. Anyway, come to think of it, it probably would be good to block them. You'd have to come up with a list of titles that wouldn't be suitable, but it would certainly work. – HDE 226868 Dec 12 '14 at 2:03
  • 2
    @JanDvorak How about saying that for people with fewer than 20 questions, you do a text similarity match on the title, and another on the body, using whatever your favourite similarity algorithm is, and reject if your new question is too close to any of your previous 20? (Picking 20 to reduce the amount of work needed on the server, and thinking that by 20 questions you've probably got the hang of the site) – Gagravarr Dec 12 '14 at 6:40

You must log in to answer this question.

Browse other questions tagged .