There are approximately 366375 observations that have zero ID applied (true “Unknowns”) AND are not labeled as captive/cultivated. (Side note, go team! That is a way lower number than a month and a half or so ago!)
Below is the data from site stats, showing the actual scope* of Unknown observations:
*I don’t know if here, Unknown means no ID at all, or includes things like kingdom disagreements and “state of matter: life”.
That’s a wide margin between 366375 and 6.2 million. How many of those are cultivated/not wild, and may be perfectly identifiable, vs. observations that simply are faulty?
It seems like we should have a better way to distinguish. While I will admit to getting grumpy sometimes and sending potted plants to iNat limbo without a label, I also get grumpy that I can’t just hit a DQA and send landscape photos to iNat limbo and have to classify them. In the iNat system, the cultivated plant can still be identified far better than a landscape photo, but they get worse treatment.
By iNat’s own terminology, I think that it’s the “verfifiable observaton” defintion which is getting abused.
“‘Verifiable Observation’ -
This status is applied to any observation that is uploaded that contains a valid date, a location, has photo or sound, and isn’t of a captive/cultivate organism. Without any one of these vital pieces of information, an observation cannot reach Research Grade and is not shared with the ALA. However, such observations can still be uploaded to iNaturalist, and they will have a ‘Casual’ status.”
This is not what “verifiable” means… Missing valid date, location, photo or sound surely makes an obsevation non-verifiable, but why does being captive/cultivated come into this picture at all?
An observation of a captive/cultivated organism that has date, location and photo/sound is not missing “any vital pieces of information” as implied above, it is still perfectly verifiable and should be treated as such.
This topic has been on my mind over and over again. This along with a comment somewhere about how “captive” examples of species could show an extension of the “wild” range map.
Could another grade be added such as city-grade?
Something like ginkgo biloba may have been planted by man, but it thrives on its own could be in this category. Then, perhaps, range maps could show wild and city-grade in different colors.
This GitHub thread from 2015 was linked recently on another thread and it contains some interesting old staff discussion about the reasoning behind the current Casual/Needs ID/RG trichotomy.
At one point they did consider having separate “Casual” and captive/cultivated:
I think there’s likely some missing context from team in-person conversations after this, as the conversation continues on a tangential subject about RG and Needs ID being mutually exclusive before settling on this trichotomy without any further discussion of why the captive/cultivated category was eliminated.
The flowcharts explain well how things are working now.
My question stems from @DianaStuder ‘s comment about how city-dwellers engaging with nature may often be not wild. At what point does something cultivated/captive become wild? And, if it is now wild, should it have its own category or forever continue as cultivated/captive?
As I’m writing this, different species come to mind that might fit as well such as wild hogs and wild horses for example.
This would address the feature request. I’d modify the first decision to include the other DQA flags (location not correct, single organism, etc.), but otherwise this would do. Provided that then the default settings on the ID modal include the “not wild” branch.
Thanks also for providing the context earlier about the communication from 2015. It helps to see that this was discussed though of course that doesn’t solve anything.
Except that many people who identify wild plants have a (in my opinion, strange) hatred of cultivated plants and would absolutely stop identifying if they could not filter them out of Needs ID.
Agreed; much as I think the core data problem of the C/C state overwriting the Needs ID state must be addressed, I think it’s important to preserve individual users’ ability to actively configure their filters as desired to obtain the results they want, and conditionally disabling certain data entry fields would undermine that.
Personally I lean towards some of the ideas explained in the flowchart above, where non-deficient observations stay in the “Needs ID” category until identified, and then are graduated either into “Not Wild” or “Research Grade” once there is a community ID.
They can hit “reviewed” and move on – same effect, but with the added benefit that observers who have engaged with nature (the primary audience) would get an answer to their question “what is this plant” sooner than later (or never).
There is a bit more on this if you scroll further.
Carrie Seltzer states:
Hi guys. I wanted to chime in because I’ve been following this and I know you’re meeting today (if you haven’t already!).
I’ve lost track of the nuances of proposed functionalities of A-D and 1-4, but I just wanted to voice two concerns of mine.
That a non-wild observation can still get “needs id” status (I know that’s one of the options outlined above, so this is my vote for that).
[…]
Loarie responds:
That would be C option 2 as opposed to C option 4. I’m personally in favor of C option 4 because IMO we shouldn’t put observations of captive organisms or observations without metadata like photos or locations up front on the site by default (e.g. people should have to go looking for them to surface them) because they distract from the core use of the site which is sharing observations with a community for ID and creating high quality data for science along the way. IMO, the additional complexity of doing this separately for captive AND obs without metadata outweighs the benefit of this very small marginal use case of people posting captive things for ID and IDers wanting to/being able to ID captive obs. But I hear both sides of the argument.
I think this is the closest we have to an official statement on why captive obs can’t have the Needs ID label.
Oh wow yeah thanks for pointing that out. Even further down from Kueda:
I think some of these decisions make sense when you have tens of thousands of observations and hundreds of users, but less so when you have hundreds of millions of observations and millions of users with different use cases. At this scale demand for more discrete categorization starts to make more sense I think, and it would have been hard for them to anticipate that at the time.
[quote is from @loarie in 2015] But all these years and … 300 million obs later, the situation needs re-evaluating. Not Wild obs Need ID. And that data is used by scientists. This is not about ‘my dog’. It is about tracking invasive sp, or phenology for blooming time for pollinators, or … That niche has exploded !!
Among the species I identify are Gallus gallus, Meleagris gallopavo, and Pavo cristatus, all of which (along with Numida, which I’ve identified at least to family from a feather) the landlady where I used to live had, and may still have, captive, and Bos taurus and Equus caballus, which I’ve observed captive. I’d identify these whether they’re captive or not. I also identify some tree species and don’t care if they’re wild or not when identifying them.
If I come across an ob of a plant in a pot with no ID, I’ll at least identify it broadly (monocot, dicot, conifer) and then mark it as captive.
1: The feature request (and from my impression also the conversation over the years) is not to treat captive and no metadata the same. It’s OK if incomplete observations get deprioritized.
2: Ten years and so many duress or CNC users later, and this is absolutely, positively, and emphatically not a very small marginal use case.
The example later in the conversation about the pet dog is cute, but misses the mark at least in my experience: most captive/cultivated observations I see are plants. @loarie, @carrieseltzer – what would it take to revive that conversation? Is there anything us outsiders can do to help?
Agreed, it is not a small use case. Most casual iNatters and the general public don’t think of street trees or park trees that got planted years before they used the park as “cultivated.” Plus, I suspect a bunch of cultivated plants don’t get marked as such because the observer didn’t mark it AND it’s not obvious enough to identifiers (and/or, we are tired and we forget to mark the DQA). So cultivated plants are going to be undercounted.
That all said, they have value from a phenology standpoint. And whether we like it or not, it’s a lot of what surrounds people, and that is nature to a lot of people. If iNat is going to trend toward being accessible to the casual user, this is an issue that is only going to become bigger.
EDIT - thinking about this more - the more urban an area is, the more likely that a person’s observations are going to be cultivated plants. A lot of humans live in urban areas. Animals are largely going to be wild regardless of where a person is, so to me, not adjusting to address the plant issue is a bit of plant blindness to me as well as blindness to how people interact with nature in cities.
It also fails to address the fact that, if iNat tilts more toward casual users, there are usually more people who misunderstand the platform as an identification app. People take photos of plants around them because they want to know what they are. Or because they are pretty (or both). But iNat isn’t a “take a photo and know what it is” app, and I don’t think it should be.
Thank you for your service! 
