As suggested when this topic was closed https://forum.inaturalist.org/t/add-observation-automatically-tagged-my-observations-with-incorrect-names-based-on-photo-filenames/8203/59, this is a Feature Request that the automatic photo file name identifications not be attributed to the observer. It seems to me just plain wrong to say that someone suggested an ID when they did not.
Personally I would rather see some kind of pop-up notice, during upload and before saving, that some IDs were automatically filled in based on photo file names, and should be reviewed and manually corrected as needed before saving.
That way users are reminded that they still have control over what IDs get uploaded, instead of allowing a bunch of nonsensical automated IDs to get into the system.
8 Likes
That would be good, too, but I don’t think everyone will understand it, especially new users. I think it probably happened here: https://www.inaturalist.org/observations/36540611#activity_comment_4117209
Imagine being a new user and getting the pop-up, and having a snail ID on a photo of a flower, and having it attributed to you, and then you have a maverick ID and don’t know what to do about that either.
3 Likes
Thing is, many of those names are by the observer. For the many, many of us who include the name of the species in the file name the ID should absolutely be attributed to the observer.
In my opinion it’s the observer/poster’s responsibility to double check any and all IDs and such upon upload and to take responsibility for correct or incorrect IDs that may be attributed to the observation at the time of upload. I very much disagree with the “upload and forget” type approach.
Check your IDs to make sure that they’re correct, it’s a simple thing to do.
4 Likes
I do. But I still think it’s wrong when someone does not do so to state that they made the identification. I think it would be better to direct the fix to people who are more familiar with how this works–like have it not attributed to the observer unless the observer affirms that all identifications on the Upload page were in fact their identifications. Or have an account setting for adopting all of these automatic observation’s as one’s own, with the default being not to adopt them. iNat is very hard to understand for a lot of people, apparently, and this thing about identifications being made from a couple of letters in an original file name was never even announced on iNat, I don’t think–or even described on iNat after the fact. I don’t think anyone not on the forum could possibly understand what’s going on.
2 Likes
Another reason I think it is wrong to attribute IDs to people who did not make them is that these weird IDs from a couple of letters in an original file name prompt a lot of snarky comments because the identifiers don’t understand the fact that the observer did not make the ID.
1 Like
@earthknight I think @sgene’s issue is that many users aren’t aware that the IDs are now being put into the uploader… it’s a change that has not been announced.
The observer IS making the ID. In the uploader the ID field is pre-populated with any usable name from the filename the user chose to upload. The user would normally go through and add IDs, check pin positions, add comments, add to projects, assign fields… and all of this before they submit the observations. The insertion of the ID is no different to the insertion of the pin location from the GPS exif information…
I agree it would have been really nice to get some pre-warning on the change though. That desire to be forewarned of changes before they impact was a big part of why I became active in the forums. I still to this day think there should be a direct message that goes out to all users immediately prior to a change, stating when it will happen and what the effect will be. At the very least it will stop the headscratching as users try to figure why things are behaving differently.
Anyone getting snarky over any ID is immediately in violation of probably the most sacrosanct rule in iNat: assume good intentions… a polite reminder might be in order if you see it happening to people over this matter :) [edit: and I know I have weaknesses in this regard too, I am working on it :) ]
But it also goes for staff and developers too… we need to assume they have good intentions in what they change and implement…
5 Likes
Well, I guess that’s a valid way to look at it–not my way, but valid nevertheless.
Yes, but it comes from frustration at seeing what appears to be something silly the observer came up with when in fact the observer did nothing other than post a photo.
It just seems like there should be some way to indicate whether or not someone actually agrees with these automatic IDs. I’ve proposed a couple of things already. Another could be not attributing them to the observer unless the observer has a minimum number of observations already, and at that magic number sending a notification that from now on all automatic IDs will be attributed to the observer.
1 Like
It is at the core of the Community ID model that identifiers are free to identify as they see fit. While we can look at a picture of a flower ID’d as a snail and draw the conclusion that WE think it is patently wrong or silly, we can’t truly know the motivation or reason behind it. It is easy to interpret the errant ID as malicious, especially if it appears to be a student that is being made to participate in something that they wouldn’t otherwise participate in, but for all we know it could be someone illustrating to someone else how the community ID model works. There are regular occurrences of IDs where taxa are similar in name but massively removed from each other (ie different kingdoms). Then there are language differences… perhaps snail is flower in some other language! I can remember being given snail vine seeds by my tutor at my hort course, and if I had managed to get them to germinate I might have made observations… and it would be an easy thing to enter snail vine, and then accidentally click on “vinegar snail” (if such a name existed)…
3 Likes
That happened on a recent observation of a Tufted Duck that I saw identified as a mushroom…apparently in Italian they have the same name.
3 Likes
If you make that Feature Request I will vote for it and abandon this one.
3 Likes
Thanks, I’ll consider that after this discussion plays out a bit more.
My idea may not be the best approach, but in some fashion, I would like to see an approach that educates the unsuspecting new user, in real time, about what information to double-check in their observations before they get saved/synced.
To me that could be a win-win if done right, cutting down on new-user (and maybe old user) frustrations, helping learning curves progress, and improving initial data quality in new observations.
2 Likes
That could even be just a single line of text immediately above the thumbnails on the uploader…
mockup here:
7 Likes
I’m not familiar with the apps. Does the Uploader appear on them too?
1 Like
Or … vote for my Feature Request to Opt out of taxon ID autofill from photo filename on upload
That gives both sides of this debate a simple solution. We already have the choice to opt out of Community Identifications, so why not the choice to opt out of photo file name identifications.
4 Likes
I think the opt-out would be a good option to have, but it will likely only be beneficial for “power users” who know what the autofill is and whether they want to use it or not.
I think a potentially more beneficial option to consider might be having the taxon ID autofill be “opt in”. A lot of folks who are renaming their photo files to have species names in them are more experienced users. This is a great option for them, and they are likely to know about the feature and be able to use it to save them time and use it with very few errors.
New users are more likely to not know about the feature/understand all the in and outs of iNat. If they see iNat suggesting a name, they may think it is right and that their opinion/ID they would have made is in error, and leave the auto-filled one. They might also be less likely to have made a correct ID in their file name in the first place (though not necessarily). As such, this feature is much more likely to lead to errors with new users (and to increase frustration for people who do lots of IDs).
Making the autofill feature opt-in could avoid these cases, but still leave this time-saving functionality available to experienced and high volume users whom it benefits most.
6 Likes
I already did!
1 Like
If you make an opt-in request I’ll vote for that too!
Well i just want to mention that all files that contain the place name: Puerto Rico are currently attributed to the owlflies of the genus Puer. i cleared the respective records for now … but i see more coming if this feature remains unchecked.
3 Likes