That’s not exactly the problem because
The problem with G. niloticus is that Aba is an English common name. So right now it’s technically working as intended – someone uploads a file with “Aba” in the file name, and the system gives an ID for Aba (G. niloticus).
Probably we should ask @kueda to hard code in another exception for Aba.
Edit: I suppose I should add that I’m assuming “Aba” is the culprit here. The rest of the common names don’t look like they would accidentally show up in file names, but I could be wrong.
Hmm, I tried uploading photos called “asa_123.jpg” and “aba_123.jpg” to the uploader, both with my account set to English and with my account set to Russian. The only one that auto-added G. niloticus as an ID was the English account with aba_123.jpg.
There is one bit of cleanup I did that may have affected the asa_123.jpg matching. Gymnarchus niloticus had an invalid scientific name of “ASA” on it, which I removed last week, since it in no way resembled a scientific name. But I didn’t touch the common names.
File names also look in scientific names for matches, so if the problem was asa and not aba, then you probably already solved it by removing asa as a scientific name.
I’ll see if we can add “asa” to our list of troublesome matches.
Except that when the feature overrides an entered ID, sometimes twice if I dont manage to get the timing right, and double check before upload that my ID was accepted and stayed put, it is a definite difference, and a difficulty.
I have been adapting by changing my file naming in some cases, but it is an extra challenge to have to avoid using a taxon name or something suggestive of a taxon name unless it is the observed species. Eg I often group-rename a dozen or so photos from a location eg “under the oak”, so I know they were all taken there, and dont name them individually as it would take unnecessary extra time to do so. Openjng the photos will always show me the taxon, whereas other data needs to be recorded. Tags are useful but not visible when scrolling through a long list of files.
I often have a similar naming protocol for things I haven’t ID’d.
It’s easy to simply group select those with a wrong name and remove the automatic name all at once, then go back with the suggested name.
The suggested names iNat comes up with should always be double checked anyway, so it really doesn’t take much or any extra time to do make the corrections as it’s something hat should be being done in the first place.
[quote=“earthknight, post:71, topic:8203”]
I always check that the ID I have selected is correct. The problem is that after checking and selecting it, it often will be replaced by the auto suggested one.
It is not easy to select those obs, if any, out of perhaps 30 observations in a bulk upload, which were named automatically and thus need renaming. I am not trying to determine how others work or how this change affects them, but I stand by my assessment of the impact of the change on my own workflow, despite having made some efforts to accommodate to change.
I’ve heard that complaint a few times and don’t get what’s going on with that. I’ve never had that happen, ever. That sounds like it’s a separate issue that’s being conflated with the auto-ID/Naming process.
The easiest thing would be to offer users an option to toggle that feature on or off as they individually desire.
In case it helps, iNaturalist won’t make any species guesses for words that are 3 characters or less, so oak.jpg shouldn’t trigger a species guess, but oaks.jpg will add one of Quercus.
I also haven’t encountered this. What we’d need to investigate are:
- filename(s) of the photos
- screenshots of what you’re seeing
- extremely detailed step by step instructions we can follow to try and replicate this on our computers
- browser and browser version
Just in case it’s relevant, there’s a thread somewhere on this forum where I was having problems when trying to duplicate an observation in order to select a second organism in the photo. I would correct the name in the duplicated observation, and the wrong name would just keep popping up again. I sort of quit doing many observations after that, so I haven’t had any recent experience with that either way.
Yes, that sounds like the same issue. I had it today. I always wonder if I hadn’t properly allowed the ID to enter before moving off, but I have seen the ID and the icon in there, and feel I am doing it the normal speed, then as I leave that obs and move to another one it reverts.
Here’s the original thread where I was discussing the problems in duplicating observations https://forum.inaturalist.org/t/incorrect-ids-added-automatically-when-duplicating-an-existing-observation/8903. At that point, I had no idea what photo file names even were, let alone that it was considered a feature not a bug to have identifications I didn’t make attributed to me on my observations. I just remembered that the quirk about the duplication process was that I would correct the ID on the original observation, but then when I duplicated the observation the new observation would have the incorrect ID on it instead of either the corrected ID I had just made on the original one or a blank (I think a blank is the usual when duplicating observations?).
I’m continuing to have problems with the word ‘bay’ in a file name even though I asked if you could add those 3 letters to the ‘naughty list’. ie, May 1, I uploaded observations with the photo file name Mercer Bay Loop and all obs were identified as Bay Laurel (Laurus nobilis).
Correct. Should be blank when you duplicate.
We now don’t attempt to match any words with 3 characters or less, so this shouldn’t be happening anymore. Let us know if it still is.
I had not seen that that had actually changed yet. Thanks!
Cheers, will do.