I’m having the same problem with location names in New Zealand that match species names.
2 Likes
Taxon ID autofill from the photo file name is a great ‘option’ for ‘power users’ and ‘experience users’ who name their files by species. However, it’s not an option, it is now the default!
The ‘opt-in’ feature you mention was implemented last year without warning, and no choice to opt out. Hence the reason other users are having problems when their photo file name is a location, or other etc, and they are getting random, wrong IDs based on the photo file name.
This wasn’t a problem before the new photo file name autofill feature was implemented, in fact, before that, I found that AI was mostly working really well for me selecting IDs based on the image, or at least making sensible suggestions which I could look in to further. And as more people upload images of nature, AI was, and is, getting smarter.
In my experience of watching users of iNat in the field who use the app and are less familiar with species, they are relying on AI to ID the observation based on the images, not a file name.
I’m guessing a lot of people when they first register on iNat, would go through the account settings and check or uncheck boxes as they registered. Of course, many possibly just glance at the info, and leave the default settings. I did! But as users become more familiar and more advanced with the nuances of iNat, they could possibly review their account settings and make changes. At that point they could choose the ‘Opt out of taxon ID autofill from photo filename’ if the feature was implemented. Or, a more savvy user, like yourself, could make a comment on an observation that had a completely wrong ID, and point them in the direction of the Opt-out feature if the reason for the strange ID was based on a photo file name. I’ve learnt so much from other users who have made informative and helpful comments on my obs about how to photograph something, which fields to use, which settings to change, etc.
We’re all about choices and opt in, opt out these days online. iNat is not much different from social media accounts you can sign up for. Have you seen how many settings facebook accounts have!
3 Likes
You have to remember that the more fine control you have over the settings, the more complicated a system will be, the more can go wrong with the system, and consequently maintenance and development effort increase dramatically.
I think there is a concerted effort to keep the volume of account settings down. There are some that I would like to see implemented, some that you would like implemented, some the developers would like implemented, some that scientists would like, and so on, and then if we add all those up it soon becomes massively complicated. I think they are taking the right approach by striving to keep it down to a minimum.
That said, if support was overwhelming for a setting to be implemented, and there was a well worded and presented argument for it, then it follows that they would be sensible to implement it!
I personally am in favour of keeping the system as simple as possible. It’s largely that formula that has been behind iNats success to date, and it would be detrimental to depart from that at this point. At least to my thinking it would be.
However, a really simple way to minimise the negative impacts of these sort of changes would be better communication. We are a community after all, and it feels a bit draconian to have change inflicted upon us without any sort of warning. If I was planning a trip and invited 3 other people, and then we got in the car to leave and found that the manufacturer had decided to make our car a 2 seater model and came and ripped out two of the seats without telling us… you can argue all day long over whether the car should be 2 or 4 seats, whether it should be left the way it was or improved and made more fun, and so on, but what I think everyone could agree on is that it should not have happened without some sort of notice… it’s embarassing to have to send those extra two guests away because we don’t have seats for them!
2 Likes
Until I read this, it hadn’t occurred to me that this is preventing AI suggestions that otherwise would be given.
1 Like
It doesn’t prevent AI suggestions. If you delete the autofilled value, and then click out of the field, then click back into the field it will give you an AI suggestion. It does make it a lot more frustrating to get to them though!
2 Likes
I think this is a good example for the proposition that people are not irresponsible but simply do not know what to make of the strange IDs that pop up: https://www.inaturalist.org/observations/39250129
2 Likes
@sgene that taxon has two really stupid common names associated with it… one is “Bd” and the other is “BSal”. I am going to look into why this taxon has those, because to me that seems really silly. I think birders have a thing where they call birds by a shortened form of their scientific name, like “TUME” for Turdus merula, and given that you can access taxa names in iNat just by entering “tu me” or “tur mer” I think these shorthands are just not justified. Especially because the extra name entries ensure that searches on other taxa end up returning these ones as well, often pushing the desired taxon further down the list.
@tiwane with vernaculars like these, no wonder everyone is complaining about the auto name fill!
[edit: oops, I should really assume good intentions! I might think they are silly, but birders probably don’t!]
[re-edit: I have since learned that “Bd” and “BSal” are valid vernacular names! They are abbreviations that have fallen into common usage.]
5 Likes
That’s right. I often have to clear out the Salamander Chytrids on my own uploads. https://forum.inaturalist.org/t/salamander-chytrid-batrachochytrium-salamandrivorans-appearing-as-an-id-on-wide-range-of-observations/10391
1 Like
I can understand the need to have shorthand means to refer to these things, especially in papers and so forth, but in the iNat context I seriously question the usefulness. The taxa can be found very easily by abbreviating the name entry.
Interestingly, the name Bd refers to a different taxon entirely, and given that no one has picked it up in the 5 months since these names were added I doubt they are actually of any real benefit to anyone!
I wasn’t really trying to raise the topic of the abbreviations/nicknames since there was already a discussion about allowing only the scientific names to be picked up in this topic https://forum.inaturalist.org/t/observations-uploading-with-random-species/8248/15, which was closed with a note that it’s working as intended. So I’m resigned to that. In this topic I was just hoping to make the confusion a little less for people by not having it look like the observer made the IDs.
3 Likes
Support overwhelming from how many people who visit the forum regularly, compared to the volume of iNat users who visit every now and then, or never? I upload plenty of obs, but don’t visit the forum as regularly as others, but suddenly there’s a new default feature that’s making my upload observation process more difficult, and took me a while and a lot of discussion on different forum threads to find out why a plant is automatically being identified as a Lion (cat species) because I happened to name my photo file ‘Lion Rock’, the location of the plant.
I agree the simple solution is better communication. But if it seems to be straightforward to implement a new default feature to ID observations from a photo file name, is it that much harder to give users the option to ‘opt out’ of that feature?
3 Likes
as much as I am for keeping things simple. I think they need to turn this new feature off, get it right, and then announce and release anew… it seems to be so problematic! I haven’t until now been aware of how much this annoys people, because I am lucky and use a Sony which does not allow me to rename photos on camera (which I used to think of as a flaw!)). If it did, I likely would do that a lot. It also only gives the option of “DSC#####.jpg” (assuming not using raw format) and so I even escape the “Bd” etc rubbish vernaculars…
3 Likes
Just a clarification that “opt-in” for a feature means that the default setting or function remains the same, and that you have to specifically change your settings to get the different functionality.
If the default is changed and the new setting is automatically applied to everyone, that is not an “opt-in” scenario.
So if there is an option for the user to choose whether they have a feature (or not), it can be opt-in or opt-out. Whether it is opt-in or opt-out depends on what the default setting (no new feature/ automatically have new feature, respectively) is for users.
1 Like
Just to be clear, I and others have this problem without doing anything at all with photo file names. Take a photo in that situation, the photo gets a seemingly random file name, and if that file name contains letters that match a taxon name, nickname, abbreviation or whatever, you have suddenly made a ridiculous identification.
1 Like
Annoys certain people. There are a lot of other people who are perfectly happy with the change and have zero problems with it.
That’s obvious, but it’s not a bad thing to explore ways to make it work better for the people who aren’t happy with it.
4 Likes
Yeah, but that’s kinda like moving to another country, wanting things to be the way I want them to be, getting those changes and then telling those that have always lived there that find those changes to be problematic to just lump it cos some of us like it this way. It adds massive insult to injury to then refer to them as “certain people”.
2 Likes
It’s absolutely good to explore ways to make it better and we should always be striving to do so.
A lot of the complaints I read don’t initially come across as that though, at least not until a ways down the comment chain when people actually start discussing approaches and alternatives.
It’s worth mentioning that this was simply an expansion of something iNat already did. If you linked images from Flickr iNat scraped the file name and applied that to the ID field. Implementing it to direct uploads was a natural extension of that existing behavior.
I’m fine with having an opt-in/opt-out option or a button that tells iNat to apply automatic ID’s using the AI instead of the file name, but it should be consistent across the board.
Oddly, even when I upload images that don’t have species names in the file name, but have other words (or even multiple species names) I never have the behavior folks are talking about. iNat either makes an automated recommendation from the image or leaves the field blank in those cases. I’ve never had it make a recommendation based on text that wasn’t the species name.
3 Likes
It seemed new to me because I’ve never used Flickr, and so I never had the identification box filled in automatically on my observations until around December 2019. Before that, when it was just done on Flickr photos, did the identifications get attributed to the observer? If so, then I guess anyone who had been adding observations from Flickr wouldn’t have noticed much of a change. But for some of the rest of us it could seem like a very sudden change.
I’m open to the alternative suggestions others have expressed. I’m especially interested in anything that would explain to inexperienced users that the identification originated from their photo’s file name and that if they don’t think it’s right they should override it with another ID if they can–such as Plant or Animal or Life even. I still do find it annoying on my own observations, but at least I understand what’s happening and can adapt to the extra keystrokes to erase the unwanted identifications before adding my own. But I do think the process should be more transparent on iNaturalist so that users who don’t understand what is going on will not be discouraged or get mocking comments.
6 Likes
Yes, they were attributed to the observer. I never linked my Facebook account to iNat (I don’t link my FB account to anything), so I don’t know if the same behavior we present when you loaded photos via that route, but if it did it from Flickr it probably did it from there too. The lack of automatic name scraping for direct uploads was the odd one out, not the norm.
The process could certainly be more transparent over all. Even if you’re active on the forum and know some of the feature requests being discussed there doesn’t seem to be much indication of which ones will actually be implemented, nor on what timescale.
I do think it’s important that the mods and developers make the changes they deem necessary with some community input, but not be constrained or 100% directed by that community input. If there was a public discussion prior to every single change nothing would ever get done.
I think it’s best of changed are made and a brief explanation of them is made to the main iNat loading page on the sidebar.
Of course, with any changes there are likely to be some unintended side-effects and those do need to be worked out, but it’s rare that any changes needs to be rolled back, and I don’t think that they should be rolled back.
4 Likes