Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 153 Stack Exchange communities.

What is meta?
Here's how it works:
  1. Any Stack Exchange user can ask a question
  2. The community provides support, votes on ideas, and reports bugs
  3. Your voice helps shape the way Stack Exchange operates

There have been discussions lately about how to improve the detection and flagging of duplicate votes. While some may be hard to find, others are quite obvious. One example of the latter is "how to send an email in android". A quick search yields many, many results (I summarize some of them in a list below as an example). These are all very similarly formatted questions that involve two key words: "send" and "email", making it very easy to parse by any robot or at the time of posting. I do understand the difficulty of implementing a failure safe mechanism to block duplicates automatically and the annoyance of more orange popups while writing a question that could contain these two words but at the same time be very well original.

Thus, I am proposing here the idea of creating a candidate duplicates list that could be accessed from the "review" tab. This would be transparent to the OP (no popups) but at the same time would facilitate the task of finding dupes. This list, together with an incentive for finding the duplicates (either a badge, rep or flag weight as has been proposed in another post) has the potential to largely improve the tagging/merging of duplicates.

(Edit: I forgot to add this part) Regarding the queries: We can propose keyword groups or phrases that are usually employed in these cases and add them to the question check at any time. A trial period should tell us whether these keywords are useful or not (i.e, too many false positives, sensitivity too low to be useful).

What do you think? Could this work? Is it feasible? Do you have a better option?

Questions sample:

About sending emails in any way:

About sending emails without using intents:

Example of a non duplicate:

(NOTE: This is not an exhaustive list of email questions in )

share|improve this question
Required reading for any meta question involving duplicates:… – Robert Harvey May 20 '11 at 17:51
@Robert, this would not be useful for those questions with one or two dupes, but for those in which you can count (at the very least) ten times the same question. Sic from the linked blog post: "Let me be clear — too much question duplication is bad. Absolutely." The above one is an example. Searching for just "i++" would give you another. – Aleadam May 20 '11 at 19:04
What's the difference between this and the "Related questions" list? Simply that it's accessible from /review? – Josh Caswell May 20 '11 at 20:14
@Josh I don't know how the algorithm is implemented in the related questions column, but I do not think it is such a simple query for keywords nor is intended to mark the duplicates. It serves a different function, through a different mechanism. – Aleadam May 20 '11 at 20:47
There's some description in a recent post by Jeff of the list that pops up while the question is being typed in. I believe that the Related Questions list is essentially the same now. The intended purpose of that list is exactly (AFAIK) to stop duplicate questions from being posted in the first place. Why wouldn't it serve to track them down afterwards? – Josh Caswell May 20 '11 at 21:16

You must log in to answer this question.

Browse other questions tagged .