36

The blog mentions the following new features in the SE Podcast #19 summary (highlighting by me):

New changes to the site are in the works. For example, no more duplicate title, and no more “Here code. You fix.” questions will be allowed.

The recent feature changes page mentions the former feature

2011-09-20: Questions with an identical title to an existing question can no longer be submitted.

but I cannot find a reference to the latter. So, how does this feature work?

(I apologize if the answer to the question can be found in the podcast itself, but I don't have speakers or headphones available at the moment.)

Improve this question
4
  • 22
    I like to imagine it's fixed by a team of Navy Seals abseiling down from a helicopter and silencing the questioner in an undisclosable but permanent manner. I suspect however that the reality is somewhat more mundane...
    – Benjol
    Commented Sep 22, 2011 at 12:14
  • 11
    Does this mean I'll have to change my policy of "I no fix. Lazy make Ug mad. Vote close."?
    – jonsca
    Commented Sep 22, 2011 at 12:27
  • 6
    @Benjol: The military term is "fast roping". en.wikipedia.org/wiki/Fast-roping
    – Engineer2021
    Commented Sep 22, 2011 at 12:35
  • 2
    Fast-roping, sometimes known as Fast Rope Insertion Extraction System (FRIES)...heh. You could say you could "fri" the question.
    – tombull89
    Commented Sep 22, 2011 at 12:40

1 Answer 1

Reset to default
31

This is something that we have been looking at, in particular for new users posting a wall of code with no context. We have added some checks to try to avoid this, such that for new users (specifically, users with low reputation), we require a nominal ratio of characters of context (i.e. non-code) per line-of-code (up to a point - obviously you can't carry that on forever, as sometimes you do need a reasonably big code sample).

Outside of that threshold, the user cannot post the question, and is asked to provide more context to explain the code. The numbers we have in place at the moment are not designed to punish / tax the asker - at current it should be exceptionally easy to pass this test as long as you have described anything about the problem - but it should cut off the WHAM! CODE! (nothing else) posts.

Note that for the purposes of this test, we do not consider the title - only the question body.

Obviously the exact numbers (and indeed algorithm) are subject to change here, as we monitor the system - but I've certainly seen plenty of new questions by new users posted since this was enabled.

Note that this check is only currently enabled on stackoverflow. It might later also serve on some other sites, but probably doesn't apply network-wide.

Improve this answer
8
  • 8
    I'm curious how this might handle the case where the question is posted with a wall of code which isn't formatted as code, and is then later edited (by the OP or by someone else) for formatting. The initial case may be something we simply can't avoid, but I wonder if this system looks at the edits.
    – David
    Commented Sep 22, 2011 at 13:45
  • 6
    @David keep in mind that this check comes directly after the "it looks like you have unformatted code" check, which applies to exactly the same users. Edits are not affected, currently.
    – Marc Gravell
    Commented Sep 22, 2011 at 13:59
  • 11
    Thanks for all your hard work. I know it is not said often, but we know it is hard to filter out the crap and remove a lot of the manual work.
    – Engineer2021
    Commented Sep 22, 2011 at 14:21
  • Can reputation play into this somehow? Some questions are perfectly justified in cutting right to the code.
    – Justin Morgan
    Commented Sep 22, 2011 at 16:51
  • My mistake, sorry. I missed that part at the beginning.
    – Justin Morgan
    Commented Sep 22, 2011 at 18:09
  • I guess that's what happened on stackoverflow.com/questions/7530345/…? The wall of code landed on jsfiddle instead of SO then. Not sure whether that was the desired effect.
    – Wladimir Palant
    Commented Sep 23, 2011 at 16:07
  • @WladimirPalant I doubt it very much - there is plenty of context there to support code. We are logging when this rule fires; I'll check the logs later.
    – Marc Gravell
    Commented Sep 23, 2011 at 16:09
  • This is great, but can we also have some sort of an absolute length check (for continuous blocks at least)? The problem is questions like this: stackoverflow.com/questions/13697209/…
    – user200500
    Commented Dec 4, 2012 at 6:32

You must log in to answer this question.

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