Add the ability for contributors to cancel their pending contributions

Maybe only items pending approval would be eligible for cancellation? Maybe newly approved items would be automatically reverted to their previous conditions if cancellation is invoked a few seconds after they go live? Conflicts? I'm not sure what conflicts there could be apart from things known as race conditions.

@adrian​

How would this work for top contributors whose submissions are accepted within minutes?

The cancellation option would be available only while the contribution is pending (regardless of how long it is pending). But once a contribution is approved, it will show up as as approved and the option to cancel it will no longer be available. 

As for contributions being approved within minutes, it depends on what type of contribution. For example, for Top Contributors, keyword contributions are (usually) approved within a few minutes, while more complex and serious edits, like name mergers, title mergers, birth and death info, edits to title release dates, edits to existing plot summaries, etc. typically take several days to approve, even for Top Contributors. And those are the types of edits that I have in mind with this proposal, since those are the ones that can lead to the biggest problems if/when they approved. At the top of this post I have linked to a bunch of threads where people have encountered these types of problems.

It seems like it would create a whole lot of conflicts.

Can you elaborate on what you mean by that? 

The idea is the exact opposite--to allow contributors to prevent conflicts that they belatedly realize might happen if their pending contribution is approved. If contributors are allowed to cancel their own edit (before it is approved) and then start over, how would that result in a conflict?

Also, sometimes when an erroneous contribution is denied by IMDb, that can make it that much harder for the same contributor to make the correct edit the next time around. Denied contributions seem to trigger an alert that can block further editing, especially by the same contributor. Allowing contributors to cancel their own erroneous contributions will prevent these scenarios as well. 

Other scenarios involve series of contributions that should be approved in multiple steps, and the contributor realizes belatedly that they submitted them in the wrong order, or they accidentally submit multiple steps before the first step is approved. Examples can include name mergers where the acting credits (character names and attributes) also need to be corrected.

I recently had something similar happen to me, and even though I saw it coming, there was nothing I could do to stop it. It's like watching a car wreck in slow motion that you caused and you can do nothing but watch it unfold, and even then you can't fix it.

Still other examples can be where two related edits are submitted to different parts of the same data (e.g. one contribution would change a release date for a movie while the other would change the type of theaters where that release occurred, such as nationwide theaters or a limited release). The system may detect these as duplicate submissions and deny them, and then also block any further edits involving the same data. 

Also, how often does one need to cancel a submission? I've made a few.

I suppose it varies based on the contributor. It has happened to me at least a few times as well. 

It seems easier to wait until they are accepted and revert them manually.

But sometimes that is not possible, as I've tried to explain with examples. And in still other cases, staff will presume that a contribution is valid, so when the same contributor wants to then reverse the same contribution after it has been approved, that may not always be allowed. Here is a recent example of that, in which an IMDb staff member expressly said in pertinent part "[a]n attempt to delete information that was originally added by the same person is always regarded with some skepticism." 

Also, why put IMDb staff through the time and trouble of reviewing a contribution even after you realize that the contribution is in error? The idea behind this proposal is to save that time and trouble for everyone involved, contributors and IMDb staff alike.

@jeorj_euler​ 

Maybe only items pending approval would be eligible for cancellation?

Yes, that is exactly what I had in mind. I provided more detail in my response to @adrian. 

Conflicts? I'm not sure what conflicts there could be

I agree that there would be no conflicts if a contributor is allowed to cancel a pending contribution. In fact, this proposal is designed to prevent conflicts. Again, more details and examples are included in my response to @adrian. 

@adrian​ 

Here is another example, also involving release attributes, where an edit that attempted to correct release information was flagged as a duplicate and denied, and at that point the Top Contributor had no recourse other than publicizing the issue in this forum.

Although this example didn't involve a situation where anyone wanted to cancel a pending contribution, it is an example of what can happen from bad data. If a contributor realizes their data is bad while the data is pending, they should be allowed to cancel the contribution before it is approved. Otherwise, we will end up with conflicts like described in this thread.

Thread by @stefano_stefano:

The Italian tile for a film is wrong/made-up and IMDb refuses to fix it. Top contributor 2017-2022 getting sick of IMDb.

@adrian​ 

I thought of yet another example: 

A few times I have accidentally submitted the same (or a similar) contribution twice before either contribution is acted upon. Only after the second contribution did I realize my mistake. In these circumstances I would like to cancel one of the contributions while they are pending, since I know that the system may auto-detect the contributions as duplicates and might "protect" the data from any further editing. Yet there is no way to do this. 

I just thought of one hypothetical scenario that might constitute a conflict, and that would be when the IMDb data editors opt not to cancel a contribution understood to be accurate, non-duplicated and very much belonging in whichever part of the website it would go live. Maybe something like that wouldn't be hard to mitigate, though. Perhaps the system would quickly purge the record of the cancelled submission. I admit that I maybe don't understand part of what adrian expressed as a concern about the hypothetical implementation of the feature request. A conflict could be any one of or a collection of many things, e.g. contradiction or paradox.

@jeorj_euler​ Can you illustrate this hypothetical conflict with some concrete examples? Because I'm still not seeing it. 

I don't know! Well, I don't honestly see how an actual paradox could emerge. Maybe a conflict on IMDb is merely an editorial disagreement. So, a man managing his wife's public image, and she is standing over his shoulder becoming angry over the information he just submitted, something like an alias-related filmography item relocation (or something similar) that nobody else bothered to sort out for the past twenty years. He tells her that she shouldn't worry, that he can fix it, if he acts "quickly enough," so he pursues a cancellation, but an IMDb data editor sees that the information accurate, non-duplicated and very much belonging, therefore blocks the cancellation. I already suggested that such a scenario could be easily mitigated against. An interesting thing to note is that the situation is probably not all that different from what happens in the absence of a submission cancellation feature. Observe please how many times people have come to the forum to request deletion of content that is understood to be public information by IMDb standards. So, to clarify, I was mainly speculating as to the meaning of "conflict" in that earlier post by me. Back to the paradox thing, I should imagine that software can be designed in such a way that it doesn't behave in a dissonant or contradictory manner, yet sometimes software developers take years to eliminate a problem, even if it manifests as a bug.

@jeorj_euler​ 

 so he pursues a cancellation, but an IMDb data editor sees that the information accurate, non-duplicated and very much belonging, therefore blocks the cancellation.

But I don't think it would work that way. If cancellations are allowed, then once a contribution is cancelled while pending it will be taken out of the queue and no IMDb data editor would ever see it.

The only possible "conflict" I can even think of would be if a specific contribution pops up in the queue for an IMDb data editor, and while the editor is reviewing and researching that contribution, the contributor cancels it during that exact brief window of time. Maybe that is what @adrian was referring to by mentioning "a whole lot of conflicts."

If that is what @adrian meant, this can be avoided too. Basically, make it where a pending contribution can only be pending in one place at a time, and it can only be cancelled while it is still in the queue.

So to summarize, a contributor would be able to cancel while the contribution is still in the queue and not yet approved (including both automatic approval and manual approval), but once it leaves the queue and is under review by IMDb (either by staff or by the review algorithm), then it can no longer be cancelled. 

Seems pretty straightforward to me. I invite @adrian to provide any concrete examples of where or how this might not work.