175

I find the grace period can lend itself to some unscrupulous behavior on popular tags. I've often seen an earlier, minimalist answer be updated within the first 5 minutes, and incorporate something mentioned in a later answer (also in its first 5 minutes, obviously) or expanded upon immensely. It's like they wanted to be "first to post" but didn't want to wait the amount of time it would take them to craft a genuine answer. (While that is the scenario where this applies most often, I don't think my request should only apply when the question is brand new. Especially if it makes the feature more complicated to implement.)

I propose that you shouldn't be willing to hit "Post Your Answer" until you think it is a valid first draft of your answer, and that very first version should be a maintained part of the answer's version history. If I post an answer and then 10 seconds later start editing it, and edit it multiple times, that should be a new grace period, and the initial answer I posted stays intact.

This kind of behavior (especially when it's intentional) is hard to detect because you have to be on the page to see the change happen.

In short:

I actually question the value of the grace period. Once you've hit the answer button for the first time, a new grace period should start, and further edits should be tracked separately from the initial submission (with their own 5-minute grace period cycles, same as today). This should eliminate garbage "first post!" answers that are edited later on purpose.

44
  • 68
    Of all the suggestions to "fix" the FGITW, this is the only one I can agree with. There should be a record of the first edition posted. Jun 28, 2012 at 15:01
  • 8
    I agree with this. Bumping shouldn't be a concern in the first 5 minutes, but if it is, it should just not occur for edits in first 5 minutes; that's separate from maintaining version history. At minimum, make the first version visible to mods even if it's not visible to others.
    – user154510
    Jun 28, 2012 at 15:07
  • 10
    Does it matter if the end result is a high quality answer? The goal of the system as a whole is to produce good answers, not to force posters to get there in "the one true way".
    – Oded
    Jun 28, 2012 at 15:16
  • 10
    I don't believe keeping the first draft would change this behaviour one whit. Regardless, a number of high rep user, myself included, do this on a regular basis (start off with a basic, good answer and immediately expand on it).
    – Oded
    Jun 28, 2012 at 15:19
  • 35
    @AaronBertrand One more thing I've seen going horribly wrong, is people flagging the placeholder answer as not an answer, moderator arriving after the full answer has been posted, and, since there aren't any evidence that the answer was not an answer initially, dismissing a valid (imho) flag. This has happened to me on ProgSE as a user (almost all of my declined flags) and as a moderator.
    – yannis
    Jun 28, 2012 at 15:21
  • 12
    Based on the upvotes I feel like I'm just not getting it, but this still seems broken. 1) Foo posts an answer. 2) Foo edits the answer -- new revision. 3) Bar posts an answer. 4) Foo copies Bar's answer within 5 minutes of their edit, and the revision gets merged Jun 28, 2012 at 15:27
  • 4
    @Oded OF course high rep user--myself included--build incremental answers. Indeed that is a positive behavior, and this suggestion will just leave a minor trail of breadcrumbs indicating that this happened. No problem there. Jun 28, 2012 at 15:36
  • 5
    Note that downvotes still work well here; if a user is FGITW'ing an answer, and you don't like the way they are approaching it, just downvote the answer. That's what I do.
    – user102937
    Jun 28, 2012 at 15:40
  • 12
    @RobertHarvey - When you DV an inadequate "stub" answer such as "answer coming" that then gets edited into shape within the 5 minute grace period what do you do then? If you leave the DV intact then surely someone will counter act the seemingly harsh down vote on an apparently fine answer thus in fact rewarding the behaviour. Jun 28, 2012 at 16:10
  • 5
    @MartinSmith: I wouldn't remove a downvote on a "stub" answer, even if they actually answer the question later. Answers gaming the system are not useful.
    – user7116
    Jun 28, 2012 at 16:40
  • 2
    @sbi: Considering the original "answer" was, "hey I've got some code somewhere, I'll look it up." I don't see that as something we'd encourage. If it were a comment instead I wouldn't have paid it any attention.
    – user7116
    Jun 28, 2012 at 16:59
  • 45
    I'd like this to apply to the question, as well as answers. Sometimes I (or others) have added perfectly good answers, but then the question has been changed within the grace period, making answers look foolish without any indication of what's happened.
    – Jon Skeet
    Jun 28, 2012 at 19:03
  • 5
    @Lamak: I'm saying that if I write an answer to a question and then the question changes significantly, it can make the answer look like it's talking nonsense.
    – Jon Skeet
    Jun 28, 2012 at 20:32
  • 6
    In case someone hasn't noticed, this question's title has been changed (not by the OP) after submission. Jun 28, 2012 at 21:53
  • 3
    Well, I'm talking about it because I came across this last night in the list of highly-voted feature-requests. Near as I can tell, you're tapping into the perpetual frustration over FGITW, but the actual feature requested here is... To kill or hobble a different feature.
    – Shog9
    Feb 20, 2013 at 18:44

10 Answers 10

31
+100

One idea would be to limit the edits in the grace period to small fixes below a certain character count. This would still allow to fix small typos and bad grammar that inevitably creeps into some posts, but would prevent the gaming of the feature you want to get rid of.

I'm not convinced yet that the grace period should actually be removed, but if it is I would strongly suggest to have some exception for very minor edits. Else the revision history gets easily cluttered with very small edits.

3
  • 9
    If someone makes a bunch of minor edits within the grace period, I'm suggesting those count as a single edit, just like they do today if they start after the 5 minutes has expired. This will not clutter the revision history except that there will be one more entry in some cases.
    – Aaron Bertrand Staff
    Jun 28, 2012 at 15:13
  • @Aaron: That we already have. What you want instead, however, is that the very first edit (the one that creates the answer) is to be treated differently from all later edits (which all come with that 5mins grace period).
    – sbi
    Jun 29, 2012 at 10:35
  • 5
    I actually really, really like the idea of only allowing very minor edits during the grace period. It's very frustrating to see a new question, post a nice, thought-out answer, and emerge from it to find that 10 one-line answers have all popped up in the mean time (so that those users can start getting upvotes immediately, then just edit their post as time allows).
    – asteri
    Dec 26, 2012 at 13:23
24

Your question sounds like the primary use case of this would be to start arguments about who stole content from whom leading to lots of angry comments and downvote wars.

It's very common for several people to come up with essentially the same answer at the same time. In popular tags it happens basically every single time an easy/common question is posted. That doesn't mean that all those people are copying each others answers, no matter if they edited their posts.

I don't see how the additional revision would lead to better answers, but I see flame wars by people who are convinced that nobody else could have come up on their own with that great answer they posted and everybody else must be copying them. But how much can you really say from a revision history that says 12:00:00: "You should use X" and then 12:00:15: "You should use X because Y. See reference Z and here is code how to do it... etc long explanation". You still have a 5min edit time window.

If someone else also posted that reason Y or reference Z you still can't tell who copied from whom. And most probably they just came up with it independently anyway, since there are only so many reasons/references that apply to a given question.

6
  • 2
    Yep, exactly. This proposal attempts to fix a "social problem" which ought not to be there in the first case by changing technical semantics which are there for a reason.
    – sbi
    Jun 28, 2012 at 22:39
  • 2
    Though maybe the question emphasizes otherwise, I'd above all say it's nice to keep track of those FGITW placeholder answers that are really meant to get into "the first slot". (And also to explain the downvotes those answers might get, but seem out of place 5 minutes later.)
    – Arjan
    Jun 29, 2012 at 5:13
  • @Arjan: But do you really want to keep track of that? Why? Isn't it that what you actually want, is to discourage the FGITW approach?
    – sbi
    Jun 29, 2012 at 10:36
  • 2
    Indeed, I do want to discourage that @sbi. And I don't see many disadvantages of not having a grace period for the very first revision. And I don't have any other solution. (Though I recall Jeff writing FGITW is not an issue anymore, I feel differently.)
    – Arjan
    Jun 29, 2012 at 10:54
  • @Arjan: I have added my thought about that to my answer just half an hour ago. Have you seen this?
    – sbi
    Jun 29, 2012 at 11:08
  • 9
    Actually, I believe this would do the opposite. It would prevent the "you stole my answer" arguments altogether, as people wouldn't be tempted to hijack other answerers. Plus, it would eliminate all the paranoia since one could just go look at the revision history. This is a win win.
    – jmort253
    Dec 31, 2012 at 7:29
19

Revising this answer in response to a lengthy discussion in comments...

This isn't really about detecting carefully-hidden plagiarism. That might happen, but it would require very careful timing and so far it hasn't been observed "in the wild". Note that real-time edit notifications make trying to pull off such a trick even more dangerous, since anyone with the page open will see exactly what you're doing...

Rather, this is intended as a rather round-about way of combating that old Fastest Gun problem:

I sometimes try and post a short correct answer as quickly as possible to kind of "mark" the question with my answer to discourage other similar answers. I then I would edit my answer to improve it, using the 5 minutes grace period to edit without it appearing as an edit (although I'll still carry on editing after then if I can improve it, I just use the grace period to edit without worrying about writing edit descriptions).

Which I find rather interesting, because... Well, because not everyone even agrees that it's a problem! But more than that, this proposal is interesting because of what it implies:

I propose that you shouldn't be willing to hit "Post Your Answer" until you think it is a valid first draft of your answer.

That's perfectly reasonable, right? Except... What does adding a new revision entry get you? Most readers don't even bother looking at the revision list; it becomes most useful when you've previously read an answer and want to see what's been updated. Kind of a small audience for changes made in the first five minutes of an answers existence... Indeed, they mostly consist of other answerers! Ah, and that's where Aaron clarified things a bit in a comment:

This is precisely why I've asked for this feature - to make people more accountable for the garbage they throw up initially.

This isn't about catching abuse, or improving the quality of answers. This is about giving folks an excuse to mete out vigilante justice to folks whose quick editing might've otherwise spared them. Frankly, I find this idea disgusting, indicative of a mindset too concerned with the game itself rather than the outcome.

I'm open to suggestions for identifying and reducing plagiarism. But the rest of this is shameful.

5
  • I've purged the comments here to make room for further discussion. Be forewarned: I will delete without reply any comment that alleges the existence of a problem without a link to an instance of it. Have a nice day!
    – Shog9
    Jan 7, 2014 at 22:18
  • 7
    Here is a case where a down-vote is locked in because the user corrected their mistake within the grace period. Their original code was wrong, they corrected it, but the down-voter didn't notice the correction until grace expired. (There were two other down-voters who did catch it within the grace period; I was one of them.) The answerer could make some other, superfluous edit now to resolve this, and then notify the down-voter that they can now change their vote, but if their original post was an official revision, they wouldn't have to do that.
    – Aaron Bertrand Staff
    Jan 15, 2014 at 19:46
  • Add a link to this report from your question, @Aaron. Since I've already answered that, I'm not going to address it here.
    – Shog9
    Jan 15, 2014 at 20:36
  • 2
    Come on Shog, that feature request is nothing at all like mine. Could you enlighten me with the exact search term I should have used to go find that first, and why I should have even done so when the link I posted above - which is exactly the kind of thing I thought you were interested in knowing about, in the context of this feature request - happened to me first-hand?
    – Aaron Bertrand Staff
    Jan 15, 2014 at 21:06
  • Extremely weary of the OT comments; I'll drop into chat and you can get it out of your system there.
    – Shog9
    Jan 15, 2014 at 21:08
15

I actually question the value of the grace period. Once you've hit the answer button, any further edits should be tracked separately from the initial submission. This should eliminate garbage "first post" answers that are edited later on purpose.

I only see why someone could consider this problem when others do that, if that someone is mostly concerned with their rep. That's fine for their personal concerns, but the site in itself shouldn't be concerned with this, the site's concern should be to produce great answers.

If a first, suboptimal answer, followed by incremental edits, leads to a great answer — what does it matter whether this incremental progress was recorded for the first 5mins or not?

IOW: I see no reason to change anything.


As an outcome of the comment discussion below, let me take up the cudgels for the current state:

The feature where changes are coalesced is there for a reason: I do not want to see individually every typo-fixing change some author does, when I look at what changed. I want to see those changes coalesced, because it's much easier to look at them that way.

If you make me look at every typo someone fixes in their answer/question individually, the very next thing I will ask for here is a feature where I can compare rev X to rev Y, where Y != X+1, because otherwise it's impossible to get an overview of what someone changed in half a dozen quick edits.

And, FWIW, I see no reason to be logically inconsistent in that regard between the first edit (the one that creates the answer/question), and the following ones. To the contrary, since a question or an answer is changing a lot more in the beginning, when the author sees all their little mistakes and inconsistencies, and when commentators point out even more of them. The beginning of an answer/question is exactly when coalescing is needed most.

So please leave this as it is. It is a helpful UX feature that would be missed. Or if you indeed must change this, then please at the least give me the option to look at the coalesced changes.


Now, if there is indeed a problem with users stealing other users' answers, and if this indeed makes those other users angry to the point where it damages the sites goal to produce outstanding answers, then let us tackle this. Allowing to view the changes uncoalesced at least for mods would probably do.

Also, if many users indeed see it as a problem that other users sneak in dummy answers first, in order to be the FGITW, then let's tackle the FGITW problem. Let's just remove the FGIW badge, or at least, change it so that it honors answers that gain 10 upvotes without being edited once (which, IMO, leaves the badge's spirit, while encouraging the exact opposite behavior).

Pushing people to aim more for the site's goal (producing outstanding answers), rather than their personal goal (accumulating outstanding rep), is a good thing.

36
  • 9
    If someone writing good content is driven off by people stealing it and getting the credit, there won't be as many great answers (I doubt this happens often, but it's certainly plausible) Jun 28, 2012 at 15:46
  • @MichaelMrozek: Stealing is orthogonal to this.
    – sbi
    Jun 28, 2012 at 15:51
  • 6
    @sbi How? It's the whole motivation for the request Jun 28, 2012 at 15:53
  • 4
    The motivation of the request is two-fold: (1) thwarting plagiarism, or at least taking a step toward making it less attractive (2) avoiding duplicate answers. This isn't about rep. I have 40K on SO, do you really think I care about 10 points that someone got because they copied my answer? Nope. I care that there are two identical answers and one was not original. The TIMESTAMP issue I mention here is because of the extra attention an answer gets because it is earlier - even though it's earlier version was crap.
    – Aaron Bertrand Staff
    Jun 28, 2012 at 15:57
  • 1
    @Aaron: I have deleted many answers of mine because someone else's was better. Or I have helped them to get improved by pointing at missing facts in comments. I lost interest in the rep game the moment I hit 10k. Now, my foremost concern is with getting good answers. I don't care how they come into existence.
    – sbi
    Jun 28, 2012 at 16:04
  • 9
    @sbi Even if it was plagiarism? And if the current system makes that easier to hide?
    – Aaron Bertrand Staff
    Jun 28, 2012 at 16:15
  • 8
    I do not believe that the end justifies the means. I do not believe that dishonesty should be rewarded.
    – Remou
    Jun 28, 2012 at 16:29
  • 2
    @JonH: I am glad you explained yourself so exhaustive that I can so well relate to your opinion.
    – sbi
    Jun 28, 2012 at 22:40
  • 5
    If someone takes the time to write an outstanding answer, taking care to test and get references, asking only acknowledgement (not rep) for this work, only to have that answer appended to the first answer, I do not think it will encourage that person to supply an outstanding answer the next time. It is against human nature to be completely selfless in the face of complete selfishness.
    – Remou
    Jun 29, 2012 at 10:34
  • 1
    @Remou: If you would be right about this (which you aren't, because even chimps show selflessness), then the whole idea of Q&A forums would be flawed, because people wouldn't give answers without getting anything back. Yet, Usenet is >20 years old, many other forums without explicit rep exist, and people get answers in SO's chat, where there is no rep to be gained from answering them.
    – sbi
    Jun 29, 2012 at 10:40
  • 5
    @sbi if you would please read over all of my comments on this page, you will realize that I do not want to track every single edit. And I will insist again that this is not about rep.
    – Aaron Bertrand Staff
    Jun 29, 2012 at 10:58
  • 7
    sbi I don't need to explain myself. Downvotes on meta mean "I don't agree with you."...and after reading your post I simply don't agree with you. You find plagirism ok, I don't.
    – JonH
    Jun 29, 2012 at 11:45
  • 1
    @sbi also, for clarity, nowhere did I ask you to compile anything. I asked you to convince me how this proposal would make your SO experience worse. You've added a lot of words to your answer, but you still haven't convinced me. Partly because it's mostly based on arguing against things that neither the proposal nor I have suggested will happen (the nonsense about reading every minor edit). Even if that were what I were after (I'll repeat again, it's not), what % of your SO time is really spent reviewing answer edits? Be honest.
    – Aaron Bertrand Staff
    Jun 29, 2012 at 12:04
  • 2
    @Aaron: What is that now, a you-haven't-read-my-comments bitch fest? I have read every single one of your dozens of comments in this fred. I am not going to re-read them again, trying to think which of them you might consider a reply to my compiled argument. If you have something to say relating to my argument, just say so. HAND.
    – sbi
    Jun 29, 2012 at 12:13
  • 5
    sbi - I think the updated title clarifies what this is really requesting. "Make first draft of a new answer part of the permanent revision history" - He's just suggesting that whatever you put as your initial answer be your first revision, always. Edits during the 5 minute interval following that are rolled up into one revision, as usual. Subsequent edits would also behave as they do now. It's just that there would be a permement record of your original answer. I hope this clears up some of the confusion =) Jun 29, 2012 at 12:32
9

This absolutely happens.

I know because I have done it. The process I use is:

  1. See a question with an obvious answer.
  2. Submit the obvious answer.
  3. Notice another answer with a different detail or a better explanation.
  4. Edit my answer before the grace period ends.
  5. Watch for things to add to my answer until the five minutes are up.

It's Stack Overflow's version of screen cheating.

Note that:

  • This only works on easy questions.
  • The problem is entirely with the game aspect of Stack Overflow. At worst, this practice will increase duplication of similar answers; it won't hurt the quality of final answers.
  • It's almost impossible to detect with certainty.

Showing all drafts won't prove anything.

Sometimes my step #3 is a bit different:

3. Notice that my answer can be improved with a different detail or a better explanation on my own and without reading other answers.

This is especially common with really easy questions since there's always more things you can say to help out someone who is pretty clueless about a technical problem. I've occasionally gone back to edit an answer moments after I left the page with a sudden insight only to see the same idea already turn up in another answer. It reminds me of the urban legend Douglas Adams used to share; I wonder if the other person thinks I'm stealing that bit of answer.

I've gone back through my answers that I might have edited in response to other answers outside of the grace period. It turns out I can't be sure which ones are examples of the concern and which ones are convergent evolution. Remember, the questions have to be pretty easy for this to even be a possible. On harder questions, you won't have several people racing for a solution. When the answers are on the obvious side, the improvements tend to be obvious too.

So even when I know I've cheated and I have the revision histories, I can't detect when I used something from another person's answer as opposed to when I've come up with something on my own. The Fastest Gun in the West strategy only works when you already know the topic inside and out. Like watching someone else's screen in a videogame, it's arguably a skill.

Gunslingers are motivated by easy kills.

I've occasionally rushed to get the first answer and used the grace period to flesh it out on sites other than SO. It absolutely helps to have the first answer. When I have a pretty good idea that I'm not the only person reading the question, I post the minimal answer quickly. If I can squeeze the updates into the five minute grace period, I will. But if it takes me six minutes, I'm not that worried about it. My motivation to answer quickly stems from the obviousness of the answer. When a question is more difficult to answer, being first doesn't much matter.

For me, if the grace period was removed, I'd be tempted to post first anyway. The only thing that might change in that case is that I might be reluctant to update my answer quickly. I'd probably work on an answer and if someone else submitted something better then my first try, I'd submit my update soon after. My goal (not just in FGIW situations) is to always have the best answer on the page. If I see someone write something better than mine, I try to improve my answer. It's good for gaming purposes and for the quality of answers on the site. (The only reason I might hold back on an edit is because of the auto-CW feature. And that wouldn't slow me down much.)

I think there is a real Fastest Gun in the West problem But I don't think this proposal will fix it. What really needs to be done is to fix the sort order of answers.

2
  • I'm not expecting it to prove anything. I just think it would make it less appealing to game the site this way (whether "game" means steal ideas from others and pass them off as your own, or simply put up a placeholder to try to get a badge, or what have you). But it's all pretty irrelevant - Shog has proclaimed repeatedly that this won't even be looked at without evidence, and the very nature of this problem makes evidence hard to come by - in fact catching someone doing this is also arguably a skill.
    – Aaron Bertrand Staff
    Jan 8, 2014 at 14:57
  • 1
    @AaronBertrand: Ah. As a person who occasionally games the system as you suggest (but not for the badge), I doubt the proposed change would slow me down. If anything, it would make my behaviour more annoying to others. Jan 8, 2014 at 16:02
4

This is only a problem for the first 0-10 minutes. Posts remain for years, at the end, the best answers will shine

Also, is not bad to have multiple similar answers, a lot of times I understand something that was already anwered but better reworded by other user.

Also, as sth said, it often happens in easy questions very regularly. And doesn't mean they stole the answers

I think is something that wouldn't hurt, but would't help too much either.

3
  • Many posts are barely even seen again after the first 0-10 minutes! There are that many of them now. Jan 6, 2014 at 19:16
  • 1
    @LightnessRacesinOrbit Who cares about what happens on those posts then? If a question doesn't have enough use to future visitors to draw views, why is it even worth considering from a policy perspective?
    – Mark Amery
    Jan 6, 2014 at 20:53
  • @Mark: I don't know. I wasn't really making any sort of on-topic point. Jan 6, 2014 at 21:04
3

So what? FGITW often amounts to 'put up a skeleton, then fill in.' How is keeping more history going to modulate this behavior? If the skeleton gets an upvote, it gets an upvote. Unless the full version is worse, there's nothing wrong with that. If the later submission is better, and gets upvotes, it gets upvotes.

If there is a FGITW problem, posting an unrecorded skeleton has nothing to do with it.

The whining complaint about FGITW amounts to the idea that voters and OPs see 'eh' fast posts, and then don't return to read superior content posted later after more effort or reflection. I don't believe it. But even if you do believe it, this won't help.

The other claim here is that this is an anti-plagiarism change, as it makes it harder for people to get away with copying. It is not. I don't care if some initial version is a copy if the eventual, long-lived, version is not.

If the eventual version copies other people, it will be there to see, without any additional edit history. If only the ephemeral initial version cribs, then so what? At worse, the person managed to grab an upvote or two from the FUITW (fastest upvoter in the west).

2
  • Conscious borrowing from other answers without attribution sounds like a vulture culture to me.
    http://upload.wikimedia.org/wikipedia/commons/thumb/4/4e/White-backed_vultures_eating_a_dead_wildebeest.JPG/320px-White-backed_vultures_eating_a_dead_wildebeest.JPG

  • Unconscious borrowing, as well as annoying repetition of the same points in the different answers even without borrowing at all, tend to degrade quality of information ("don't repeat yourself").

Whatever stands in the way of above, I am all for that.

Being able to see first draft is particularly appealing to me because it helps me figure which answers to downvote for useless repeating (no matter borrowed or not) of what was already stated.

For the sake of completeness, I wouldn't downvote if the repeated point is presented substantially better than in earlier answer.


Proposed feature is "incomplete" in the sense that it won't help me determine repetitions that occur in grace period after first draft but I don't mind. I am not fond of splitting hairs in cases when answerer doesn't even bother to make substantial points into first draft.

1

While tracking even the first edit is nice (doesn't hurt the initial purpose of the grace period), how would it impact the behavior of those who continually edit? If their answer is good, then why would any sane moderator delete/rollback the answer?

5
  • 1
    Continuous edits would still benefit from the 5-minute grace period that starts once they perform their first edit. What I'm trying to discourage is people posting a simple answer like @Yannis did below, then come back after they've done research on it. I'm not sure whether the motivation is similar to the "first post!" mentality, or if they're after the enlightened badge, or just after rep based on time and hope people won't notice that they changed their answer based on other answers/research/etc. within the grace period.
    – Aaron Bertrand Staff
    Jun 28, 2012 at 15:12
  • @AaronBertrand For the record, the full version of my answer was very close to Mad Scientist's, since he got to the actual answer first, no point in having a dupe answer.
    – yannis
    Jun 28, 2012 at 15:16
  • 10
    No one's asking for deletes/rollbacks. This is about accountability so people can't copy other's code and say "I didn't copy, I posted first, see!" Unlike most FGITW this does not discourage answering quickly, it just holds you accountable for posting "Wait lemme just reserve this timestamp" stuff
    – Ben Brocka
    Jun 28, 2012 at 15:18
  • 2
    Ditto what @Ben said - I'm not asking about rollbacks/deletes. I'm just asking for the initial answer to be a permanent record.
    – Aaron Bertrand Staff
    Jun 28, 2012 at 15:19
  • @BenBrocka Ah, understood. Thanks for explaining. Jun 28, 2012 at 15:42
1

The original purpose for the grace period was because a certain number of edits by the OP would cause the post to become Community Wiki (currently 10 as per the FAQ). Since a lot of us make mistakes and/or remember more information to put in, they put some slack in the system to help us out.

What I think would help that this question brings up, is for all of those edits to show in the history, even if they don't count for CW flipping. You could also disable rollback for those edits within the grace period.

I understand this will be a little complicated based on the probable database/engine design, but it would cover a lot of the concerns.

Another simpler alternative would be to just disable CW flipping for editing. Then you wouldn't need a grace period at all. CW has changed a lot, so it doesn't really seem necessary at this point.

6
  • So far as I recall, the grace period predates the CW autoconversion; heck, the concept predates SO itself by quite a bit. Heck, there's a grace period for comments as well...
    – Shog9
    Feb 20, 2013 at 9:11
  • My memory was that CW autoconversion was there almost from the start, if not the start, but my memory is quite imperfect. I remember when the grace period came along, but don't have a clear memory on when the CW autoconversion was happening. Feb 20, 2013 at 15:22
  • We've .. changed the scope of community wiki. It's no longer automatic for crossing an edit threshold. I can't remember any other practical reason (or use case) envisioned for the grace period. Lance - sanity check?
    – user50049
    May 12, 2015 at 16:37
  • @TimPost, I can't think of any circumstance where it's needed. It may be useful to keep from cluttering the edit history, since my workflow (unfortunately) involves making a lot of small edits after I finish and read it some more. However, I think it would be ok to not have the grace period. May 12, 2015 at 16:46
  • Well, we could still have the grace period .. I have to think about it a bit more. Thanks Lance, as always.
    – user50049
    May 12, 2015 at 17:39
  • @TimPost One practical reason is to have readable revision history. If each "oops, missed yet another comma" edit generates a new revision, the first five minutes of a post's life would take up a sizable chunk of the history, without adding any information. The size of PostHistory table is also something to consider.
    – user259867
    May 12, 2015 at 22:56

You must log in to answer this question.

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