Change account deletion functionality to allow account anonymisation and prevent deletion of IDs

I think the discussion here mostly relates to ID’s and the expertise of the now gone user. I think the Observations can almost be treated as a separate case.

In either case - it would be good to understand why people want to delete their accounts, maybe there could be a questionnaire for them to complete before the account is deleted and the responses could be collected and analysed before any changes are made?

My assumption is that people would be less concerned about their ID’s being left, but more concerned over the observations as this can be used to obtain location/history of the person in real life.

7 Likes

See tiwane’s new Help
https://help.inaturalist.org/en/support/home
In the 4 most popular articles is - How can I delete my account?
It is tiwane’s deliberate choice to make that option clearly available for iNatters who need it. (In fact it was his first choice when he launched the New Help)

I think many people here approach this problem from an unrealistic point of of view - people are writing about how “anonymous IDs are less valuable” etc. - this is all nice and cool, but this would only be relevant in a situation, where we would be having a choice of different of quality.

But that’s NOT iNat right now. We are absolutely desperately starved for IDs - the lack of IDers is the one big glaring problem of iNat which completely dwarfs all these tiny concerns that people are coming up with here. A prolific IDer deleting their account today is a disaster. No, it’s not a “minor inconvenience”. There are people who could literally shave several per cent off my life list with a single click and I can do absolutely nothing against it as it stands right now.

For me, the problem that “IDs will be now anonymous and we will not know whether we can trust the IDs and maybe there would be some complications with how the IDs are counted etc…” are quite funny compared to the very real problem that if this one guy deletes his account, I will never again know what this was. Old un-IDed insect observations remain un-IDed forever, 100%, no hope.

11 Likes

I agree that a soft(er) delete option should be available, but it should resemble OP’s “Also, if a scorched earth option must be included…” concession paragraph.

In my ideal soft delete:

Observations should be deleted (but not in a manner that results in broken URLs). IDs should be withdrawn and attributed to “deleted account”. Comments should be deleted except for some placeholder note like “comment by deleted account”. I think this satisfies @cthawley’s very good points above about the problem of anonymity in scientific records. The deleted user’s contributions are now merely a suggestion of a suggestion, and at that point I think it’s okay for them to be anonymous, much like a scrawled note in the margin of a herbarium specimen.

Users impacted by the removal of the now-deleted account’s IDs should get a single notification that goes to a page with links to every impacted observation. This should be very similar to what is already done with the taxon change alert pages already shown to users who have opted-out of automatically accepting taxon changes. The existence of this notification needs to be communicated to the exiting user.

This places the onus of tediously fixing the missing IDs on people interested in (or even entertained by!) doing so (rather than on the likely-distressed exiting user, which is accidentally cruel), and alerts exactly those interested people to the problem. I think this harder version of a soft delete would satisfy more exiting users than the softer soft-deletes suggested above, which frankly would retain so much information that I think a distressed exiting user would be unlikely to choose them, opting instead for a total delete.

And yes, a total delete should always be an option. OP has restricted this conversation from discussing why.

9 Likes

The worst thing about the current situation is that unless you follow this forum at least occasionally, you not only do not know what can happen you, but, even worse, it may already have happened to you and you don’t know that it did! The fact that your observations can lose their ID in the blink of an eye is never communicated to you, at any stage of using iNaturalist itself.

Seriously, at this point I am considering just going through all my observations and blindly agreeing with the current consensus, just to make sure it stays recorded. It’s absolutely against the rules (because I would be making IDs that I have zero idea about), but it seems to me to be the only way to keep the information safe for now.

iNat loads an observation in ~3 seconds by my measurments, 1 more second to click, that’s about 42 hours of work, better to get started …

2 Likes

which is why ‘we ID for others’ ? Yes ?

2 Likes

I have a huge amount of respect for IDers. I have tried IDing, but found quickly, that I simply do not have the confidence in me to ID anything when I have even the slightest doubt. Sure, I could ID some easier birds in WP, but those are usually IDed very quickly by others. Then, once it’s something less familiar, I can take out a book from my collection … but then the doubt starts - do I really understand all the variability in birds that I have never seen?

Besides birds, I am really into mammals, but I find it extremely difficult to give confident IDs for those - even though I think I know the mammal, it happens so often, that people convince me wrong about my own observations. Because a lot of mammal ID is just based on general shape and feeling - I know one of the best bat experts around and he told me “I can’t explain it, but once you hold your 10.000th bat in hand, you will kinda get it”. So how do people actually find the confidence to ID, I don’t know, but those people are really critical to making iNat the thing that it is now.

The problem is really that those people are too few, especially for the less popular taxa, such as any insect that’s not butterflies or dragonflies. And now if you are so lucky and some of these experts gets to your observation and IDs something unusual, it’s a cause for celebration - but should they decide to delete the IDs, nobody will ever look at the observation again, the IDers clearly just barely keep up with new observations, almost nobody is going back in time, so any such observation is lost forever.

2 Likes

If you try identifying - and yes start with where you are confident. Pick thru Unknowns for birds. Pick thru local Aves for the ones you know. Or annotate bird obs, and learn as you work thru a batch. Even the butterflies and dragonflies - can be picked out of Unknown, or Insect, or Wings.

We do - slowly - get new identifiers. Taxon specialists do taxon sweeps. iNat IDs are not carved in stone - taxonomy changes force specialists to sweep thru, again. Biggest recent change for me was @chemp and African Orthoptera.

If you follow your notifications, you can withdraw or reconsider if someone disagrees with your ID. Choose a small and focused slice, and whittle it down steadily.

https://forum.inaturalist.org/t/how-to-best-contribute-to-identifications-looking-for-advice/39559/96

1 Like

Yes! I was shocked to learn that this is what happens!
Now I need to make sure that the list of passwords I am leaving behind for posterity does not include my iNat PW.
Or at least, I need to include a note saying “please do not delete this account.”
Because while I have not made any spectacular observations, I have at least started doing more IDing lately.

3 Likes

I want to thank everyone for providing insightful comments to this diffcult topic. I’m personally for an anonymization option, and I think I currently fall on making anonymized IDs not active side, for much of the same reasons @spiphany has laid out. But there needs to be a lot of thought put into what ever is implemented.

FWIW, Wikipedia has a Wikibreak enforcer, for people who feel like they need to take a break but also want a way to keep them off of Wikipedia for the intented period. That might be an interesting option as well.

Not an engineer, but @-mentions are not recorded in an specific way. A notification is generated, but after that it’s just a piece of text that the site/apps interpret in a robotic way - making a link out of the text and linking it to /users/[username]. If anonymization included @-mentions, I suspect there would have to be some sort of find and replace script, which would probably be super inefficient.

To be clear, it’s our developers who do this, I’m just the person who relays the request. Bringing back an account is way beyond my technical capabilities.

The only time this is really feasible is before the previous week’s backup from our test server is replaced, which happens every Saturday, US time. For example if an account is deleted on Friday and we don’t get the request to resurrect it until the following Monday, it’s too late - the backup that had their easily-accessible information is gone.

10 Likes

Should be able to do something similar to Reddit where the comments from deleted profiles remain in a thread, but the original profile name is replaced with “[Deleted]”.

4 Likes

Alternatives include:

  • Leaving a comment on the observation stating the Research Grade identification.
  • Editing the observation with the Research Grade identification stated in the Notes section.
  • Employing an Observation Field for “Research Grade at:” with the Research Grade name filled in per observation.
  • Agreeing with the Research Grade identification but then withdrawing your ID so the name is recorded without you asserting that you agree and can confirm.

Edit: I used ‘Research Grade’ in my examples but the same idea holds true for any identification someone has made.

I just wish it were easier to recognize these from the Dashboard page. Not the most convenient thing to browse through 300+ entries and open the ones I think might be a disagreement to make sure I’m not missing any. I imagine it must be more time-consuming for users like you who do even more identifications.

Would be great to see this addressed in the Dashboard rework.

7 Likes

If the IDs become inactive, it’s really important that the observers get notified about it clearly - otherwise it almost doesn’t matter that it wasn’t deleted.

It’s the same as with the idea of “marking doen the ID in another way” - in principle yes, it saves the information, but if you don’t know, that the ID is lost, the observation is now still not IDed.

2 Likes

i tend to agree with this general sentiment. i think it’s fine for a deleted user’s identifications to be deleted, and i would just have the system add a comment only in cases where the deletion results in a change to the observation taxon, to note what the observation taxon was before deletion. this will allow folks to be notified that the observation id changed and take action, if needed.

i don’t think i would make account deletion more complicated, nor would i try to allow a deleted account to leave some of their contributions behind, even if anonymized. instead, i think i would provide users an option to simply deactivate their account. this would add an indicator on the user’s profile that they are no longer active on the system and would disable all communications channels to the user. the user would have the option to reactivate their account, but they would have to go through some sort of process (similar to a password reset) that goes through their e-mail address on record.

5 Likes

to be honest - when in doubt - I click to open. I need to see WHO the identifier is. If the IDs are anonymised - they should not count towards the CID - as if they were ‘mentioned in a comment’. If my ID was cautious I want to see if the support is informed or just polite.

Offering a time out, cool down, option would surely prevent some account deletions.

2 Likes

That’s a good point. I learned in other fora that some people get “addicted” to the forum, and spend more time on it than thy can justify. And then opted for the deletion of their account. In one extreme case, the user even asked the admins to prevent him from creating a fresh account…
But for most of such addiction cases, a “prevent me from using the site for the next … weeks/month” should do.

But let’s look at a different point: protection. For what ever reason, a user may feel in danger due to his use of iNat. It does not matter if the reason is real, only that the user thinks so.
Imagine that you (yes:you yourself) spent all your time in the office with iNat - uploading yesterday’s / last weekend’s observations, identifying other observations. From Monday through Friday from 8 am to 5 pm. And now you think your bad colleague detected it, and might tell it to Big Boss.
How can you protect yourself?
Change your username? It’s not your real life name anyway. That would not help much.
Anonymization such that every item gets a different (non-predictable) username? But what about the time stamps of observations, upload time, comments, ids, …? And the places of observations? Would someone else still be able to associate that with you?
Better get rid of everything, and thus all traces are gone, and see: that colleague was just lying?

Now you could think that such behavior is bad, and hence protection not adequate. Hm. What about iNat becoming illegal in your jurisdiction? That’s not a crazy idea of mine. Antivirus programs of Kaspersky are just becoming illegal in the US, and even as a private user of them you may face prosecution in the future. After all, we live in a crazy world.

3 Likes

I know this is just an example you made up, but If there was a button that you could click that automatically changed your username, wiped your profile and picture and unsubscribed from email (all things you can already do), what more would there be to prove that you are connected to your account? Yeah that user IDed lots of beetles and your coworker knows you are good at beetles but so are lots of other people. how is he supposed to prove that was you. Unless the coworker is a close friend that knows a lot about you (in which why is he ratting you out?) time stamps and locations are awfully circumstantial and not very hard evidence (especially if obscured)

In this case, after you clicked the button, you still know what your password was and might have memorized the auto generated username so can regain access if needed. that’s not the point of the button of course, it’s simply to grant the ability leave the community but and not be able to be traced by the general public back to you personally (at least make it insanely harder)

Just because you already have the options to do all those things manually doesn’t mean you would realize that by doing all those things, you would be sufficiently untraceable by all but staff and yourself. Right now you’re afraid that at any moment you might be on your way to being fired and deleting is the safest bet you see. If there was a second button with a description like “you and your personal not be traceable by the general public ” I’m sure it be used just as much if not more than the delete button. But I suppose for that to be really ba a true statement it’d have to delete your comments because you might have mentioned in a comment that trip to Alaska where you caught 11 pike and your coworker also knows about that story. Your coworker is probably not going to dig through the entirety of an account he’s only mostly sure is yours for some scrap of personal info like that but you’re afraid and want to be able to sleep at night so you hit that delete button.

Or not. Maybe you’re not that afraid but still feel you need to do something as a precaution. Having that button (even though it doesn’t do anything you can’t already do) makes it obvious that there are other options that are almost just as good but doesn’t destroy everything you worked so hard to find because chances are you really don’t want to have to delete your account.

Yeah someone could theoretically figure out it was you (depending on what kind of stuff you said in your comments) but you really don’t want to delete your account. I’m not saying to remove the delete button, I’m saying add a new one. You’re coworker is being paid to work, not get you fired, he’s probably not going to try THAT HARD, and probably doesn’t know how to navigate iNat anyway. Nonetheless the delete button is still there.

Addressing this item and more. Allow for observations and other data such as annotations to be donated to some institution(s)/organization(s) (this could include iNat). It is like when one donates a physical specimen to a physical collection or makes a physical annotation to a specimen housed at collection. Those observations and annotations are no longer owned by the collector or annotator but by the institution/organization. If this was an option on iNat, all data that was donated could be archived indefinitely. People could be encouraged to donate their data so that it would have more of a chance of being archived into the future.

It is important to think about this data into the future. After people die and others become in charge of it. If data could be donated then it will have more of a certain future.

For those not interested in donating their data there is still a need for what is being discussed. But having a donation option would help in some situations.

7 Likes

I agree with your argument but also just want to provide an example from the other side of this. In some cases it’s good that all the identifications get deleted because some people aren’t acting in good faith. A couple years ago a user went on a spree of providing incorrect, higher-level identifications on many of my observations and on many other observations in my area (think disagreeing with a butterfly identified to species and marking it as being a beetle or even a plant). The nature of how incorrect many of these higher-level IDs was seems like this was intentional on their part. I assume their account was deleted by admins after being flagged, but in that case I was very glad all their IDs disappeared because they were making a lot of extra work for a lot of people and it was really frustrating. While this probably isn’t nearly as common of a scenario as someone who deletes their account and had identified in good faith, it still would be good to retain the ability for an admin to remove all a user’s activity if they are intentionally trying to disrupt the identification process.

6 Likes