4
$\begingroup$

As I just found out, multiple people can edit a question at the same time.

I happened to be the one to submit last. The red notification that told me that I cannot submit that edit was really pretty, but I have to wonder why the system allows many parallel edits if it's not capable of merging the results. Why does it bite more off than it can chew?

I don't feel like wasting more time today trying to submit edits.

additional information:

  • I have less than 2000 reputation.
  • I click on "edit" below whatever I want to edit in order to (try to) edit it. I apply changes to the text.
  • I hit the submit button, which is refused, because there's another edit waiting to be reviewed.
$\endgroup$
  • $\begingroup$ Maybe relevant: meta.math.stackexchange.com/questions/8917/… $\endgroup$ – Thomas Sep 4 '15 at 13:05
  • 1
    $\begingroup$ You should likely make precise if you are taking about "editing" or "suggesting an edit" (or both). $\endgroup$ – quid Sep 4 '15 at 13:32
  • $\begingroup$ @quid I do not have more than 2000 reputation. I am suggesting edits. I cannot tell if this is also about regular edits. $\endgroup$ – null Sep 4 '15 at 13:55
  • $\begingroup$ Yes, this is what I thought (at least on second thought). Note though that you can edit CW posts and your own posts, so it is not strictly true that you can only suggest edits. But the problem is present too for regular edits. $\endgroup$ – quid Sep 4 '15 at 14:00
  • 1
    $\begingroup$ This StackExchange's homage to the Highlander franchise. $\endgroup$ – Asaf Karagila Sep 4 '15 at 15:29
  • $\begingroup$ @asaf, it turns out that this also happens with species where multiple males copulate with a female over a relatively brief time. It was, supposedly, an undergraduate biology major who pointed out that this meant sperm competition continued after copulation. press.princeton.edu/titles/7232.html $\endgroup$ – Will Jagy Sep 4 '15 at 16:49
  • $\begingroup$ @Will: I actually had this in mind... $\endgroup$ – Asaf Karagila Sep 4 '15 at 17:42
  • $\begingroup$ @asaf, I enjoyed the highlander TV show. Not sure I saw the movies $\endgroup$ – Will Jagy Sep 4 '15 at 17:47
  • $\begingroup$ meanwhile, while some undergrad may have found something separately, the consensus is that the idea originated in this: onlinelibrary.wiley.com/doi/10.1111/j.1469-185X.1970.tb01176.x/… $\endgroup$ – Will Jagy Sep 4 '15 at 17:49
  • $\begingroup$ At the point you are notified the edit cannot be submitted, the edited text is still on the screen and, if you are intent on submitting it, you can copy it for later submission. Of course this involves some deferred gratification, but the system is not designed to merge edits. Someone always has had the last word in the edit history (even if the last word was rolling back someone else's edit). $\endgroup$ – hardmath Sep 4 '15 at 18:33
8
$\begingroup$

It is not easy to know for the system if/when somebody is actually editing the question. It would likely be possible to prevent all users from starting an edit once one user has clicked "edit" but for how long should one maintain this if no submission is forthcoming?

It is not rare that a user clicks "edit" and does in fact not submit an edit (for one reason or another). At least I do this with some frequency.

However, if there is an actual pending suggested edit this blocks further submission. So it is not that nothing is done, it is just that there are some trade-offs.

Also note that the system in fact tries to decide if an edit is more substantive than the earlier one, and would apply it if it thinks it is.

$\endgroup$
  • 1
    $\begingroup$ +1 for sharing the process behind it. But your answer implies that there's a decision between several edits. Is there any effort to merge both? If not, why not? $\endgroup$ – null Sep 4 '15 at 13:57
  • $\begingroup$ I am not sure. But it could be complicated to avoid conflicts. Even if say edit 1 was only on paragraph 1 and edit 2 on paragraph 2, the former could invalidate the latter. Now, this might be rare, but one might prefer to be conservative. $\endgroup$ – quid Sep 4 '15 at 14:05
  • $\begingroup$ IIRC the software blocks competing edits in certain review queues (or perhaps in all of them). This is practical because within a review, there is usually an explicit signal "user abandoned the edit", namely the user clicking Skip or picking another option. (If the reviewer simply navigates away from review, leaving the task unfinished, the lock persists for some time, ~15 (?) minutes.) $\endgroup$ – user147263 Sep 4 '15 at 17:15
  • 1
    $\begingroup$ @null attempting to automatically merge multiple users edits to a single file isn't uncommon in software development. As long as the editors are working in different parts of the file it mostly just works. When it doesn't however it tends to fail messily. Some of the failures are detected automatically and trigger a prompt to the user that they need to fix something; other times the tool produces a false negative where it thinks it was able to merge two edits successfully but ended up screwing them up badly. I assume SE decided the risk of bad merges was too high to allow them to happen. $\endgroup$ – Dan Neely Sep 15 '15 at 15:07

You must log in to answer this question.

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