69

Sometimes (eg. this question) more than one user will simultaneously attempt to edit a post, particularly to fix poor formatting by the OP. It looks like in that case, I finished 30 seconds sooner than @marcog, but I don't know who might have clicked "edit" first.

Anyway, a useful usability enhancement would be to keep track of who is (or might be) editing a post at any given time. With this enhancement, the second user to click "edit" might see something like:

The following users may currently be editing this post:

  • marcog (21 seconds ago)

If a message like that were available, I would immediately choose not to edit at that time, and maybe come back later if the formatting still needed fixing. This saves effort, frustration, and duplicate "hey I fixed your formatting" comments.

It is important to note that this feature would not attempt to prevent multiple editors, everything else would work just the same as it does now.

Implementation idea

This feature could be implemented by adding a new table, say "editors", with the columns (post_id, edit_start, user_id). When a user clicks "edit" on a post:

  • Check the editors table for old entries for this post, say > 5 minutes. Delete them.
  • If there are any current editors, show the list as above.
  • Insert a row for the current post and the current user with the current time.

When a user submits changes to a post:

  • Delete the user's record from the editors table for that post.

Periodically:

  • Scrub old entries from this table (if a user clicked "edit" but never submitted a change for that post).

Optionally:

  • Provide a "Cancel" button on the edit page. Cancelling would remove the current user's record for the current post, but obviously not submit any changes.

Update: After saying in the comments that I've never seen the edit notification, I actually did just get it today editing this question. However, it showed the orange bar at the top (1 other users edited!) for a brief time just after I pressed "Save". What good is that?

17
  • 5
    The system already warns you if someone completes an edit while you are still editing. Isn't that enough?
    – ChrisF Mod
    Commented Jan 4, 2011 at 21:42
  • 2
    Do you notify all editors? If so, won't they all hold back their edits? (Sorry about the edit clash though! Although no-one can be blamed, I always feel bad when it happens.)
    – moinudin
    Commented Jan 4, 2011 at 21:44
  • @ChrisF What happens if both of us were doing non-trivial edits?
    – moinudin
    Commented Jan 4, 2011 at 21:45
  • @ChrisF: I've never seen that warning. I've seen the one where you get notified as more answers appear while you're writing an answer, but not for editing. Commented Jan 4, 2011 at 21:45
  • @marcog - if you get the notification you can abandon your edit, check the current state of the post and re-edit if there's still something wrong.
    – ChrisF Mod
    Commented Jan 4, 2011 at 21:46
  • 2
    @ChrisF The case here: A starts editing 12:00, B starts editing 12:01, A finishes editing 12:06, B finishes editing 12:06:30. a) no warnings, and b) even if the warning were to appear, B has wasted 5 minutes.
    – moinudin
    Commented Jan 4, 2011 at 21:49
  • 2
    @Greg I've seen it, but it only checks every 45 seconds and since there were 30 seconds between us finishing our edits it never triggered.
    – moinudin
    Commented Jan 4, 2011 at 21:50
  • Indeed, my formatting cleanup edits rarely take as long as 45 seconds. Commented Jan 4, 2011 at 21:51
  • @marcog - unless it checks constantly you are always going to get clashes. I don't know what the answer is.
    – ChrisF Mod
    Commented Jan 4, 2011 at 21:52
  • @ChrisF: Some edits could take time, and alerts when finishing are too late. Maybe it could check and emit alerts when saving a draft... but I don't if question editions have drafts.
    – user150068
    Commented Jan 4, 2011 at 21:54
  • 1
    @ChrisF Ideally the concurrent edit heartbeat would be more frequent for the first minute of editing, but informing people up front of multiple editors seems like a more reliable way to reduce this issue anyway.
    – Tim Stone
    Commented Jan 4, 2011 at 21:57
  • @Tim Ideally there would be some sort of collision merging interface, but I don't have high hopes for that :) Commented Jan 4, 2011 at 22:03
  • 5
    @Michael Mrozek: When Wikipedia/MediaWiki can't do a 3-way merge, it allows the editor involved in an "edit conflict", who attempted to save later, the option to retry his edit, but not in an elegant fashion. He gets the new version, his version, and the diff between them back, but no more. However, our current system is just plain broken: it alerts an editor via AJAX to others who saved edits while he was editing but does not even show the changes made. Commented Jan 5, 2011 at 2:18
  • It's rare, but some edits take more than five minutes to complete. (see revision 16)
    – Pops
    Commented Jan 5, 2011 at 18:36
  • 2
    I like this, we track this info already with the drafts feature ... however I think Jeff did not feel that strongly about it
    – waffles
    Commented May 3, 2011 at 6:37

2 Answers 2

1

I am not in favor of this, because it will create a chilling effect on edits.

That is, just because someone else is editing, doesn't mean

  • their edit will be any good
  • their edit will be more substantive than yours

.. and, besides, we already have multiple mechanisms in place to deal with this.

The risk of even less editing is far more dangerous than any minor infoporn benefit this would have.

7
  • 3
    (I might just be very slow on the uptake, but) Could you define "infoporn"? Or is that just a typo? Commented Jan 13, 2012 at 16:34
  • 2
    Fair enough, but wouldn't that be a decision the second user (the potential editor) could/should make? Sure, the other edit could be less substantive. But "our" time has its value/cost too? "We" could be spending that time on answering/editing another post, while the first editor might very well be making the same edits "we" want to do. (Besides: I think the current mechanisms are simply not as good as they used to be!)
    – Arjan
    Commented Jan 13, 2012 at 18:05
  • (Another aside: it seems I can award the bounty to the deleted answer from Balaswamy. I could try, if you want me, if you promise me I can revert it...)
    – Arjan
    Commented Jan 13, 2012 at 18:08
  • (Just to be sure: it seems you refunded the bounty, but I did NOT try to award it to Balaswamy answer. I don't mind "spoiling" a bounty, I just didn't want to award to some random post.)
    – Arjan
    Commented Jan 13, 2012 at 19:58
  • @arjan right I assumed you didn't want to award it to this post, so.. :) Commented Jan 13, 2012 at 20:20
  • (I would happily have awarded it, as the community seems to agree with you, and even [status-declined] is an answer! ;-))
    – Arjan
    Commented Jan 13, 2012 at 20:30
  • 2
    Jeff. I see the following as a perfect analogy. If you sat down to write something on your computer and was sometimes met with a message saying that there is 50% risk of you loosing everything you write for the next 5 minutes. Would getting this message make you use your computer less for typing in general, compared with a user who never gets this message, but occasionally looses something he wrote. I think it is obvious that the latter would always think twice before typing something on the computer.
    – David
    Commented Feb 13, 2012 at 2:50
1

From Jeffs answer:

I am not in favor of this, because it will create a chilling effect on edits.

Indeed, and sometimes that is good. We have the same effect already after you approved a suggested edit. You have to wait minutes on SO, sometimes hours on other sites, before you can edit again. This doesn't seem to be a valid reason to me.

On active sites, with a lot of experts in a tag, it happens on a daily basis that two people edit the same question, fixing the exact same code, indentation and spelling. It is a waste of time of those users to let two or three users edit the same post. If I can see Jon Skeet is editing the post, I will be happy with it. I know he does a good job at that.

Please reconsider this functionality!

5
  • @Sha: sometimes English is hard... I didn't fully recover from my night sleep yet. Commented Jun 6, 2016 at 6:14
  • lol, no worry... it was funny. ;) Anyway, while I kind of agree, answer on a declined request isn't likely to draw the required attention. IMO in those cases, it's valid to post a new feature request, making it crystal clear you are well aware it was suggested and declined before, and making new cases why the team should reconsider. Note that others might disagree and close it as duplicate anyway. Commented Jun 6, 2016 at 6:15
  • Hmm, okay. Won't that pile discussion on discussion? @Sha Commented Jun 6, 2016 at 6:16
  • Well, the second option is editing the existing request with new reasons, but again - I'm not sure the team track requests or bugs that are already declined (or completed.). This is rare enough to not cause problems, sometimes there's simply no better choice. :) Commented Jun 6, 2016 at 6:18
  • Okay, will try to work on a new request. Commented Jun 6, 2016 at 6:18

You must log in to answer this question.

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