What about something like a “Hot! 
” label, attached to any observation with non-agreeing interactions above a certain threshold/over a certain duration? (i.e. comments + disagreements + CID reversals + taxonomy changes and whatnot)
Sure, there are probably some observations that would fit that scenario.
But imagine a scenario where several taxon experts have looked at an observation and given it the same genus level ID because they are stumped. It’s a tough ID, but they want to leave it at Needs_ID because maybe somebody else will come along with additional insight. In this scenario, there might be no disagreements or CID reversals. There may be a few comments outlining the problem, but there may not be enough to trip the “Hot!” label.
What we don’t want to happen is to have someone with little to no expertise with the taxon coming along and casually applying a species level ID “just to get it out of the needs_id queue”.
In my experience, variations on this scenario happen a lot - even in cases where there was a long discussion and explanations for why the ID may have to remain at the higher level. My guess is that the hit-and-run IDer entered their species level ID from the thumbnail view, and never even saw the comments. It forces me to go back and re-apply the same higher level ID (rinse and repeat). Instead of have the flag triggered by a comments, disagreements, and CID reversals that may or may not happen, why not have a flag that can be upvoted by the folks who looked at the observation in detail.
From my point of view, this is the more problematic scenario than the one your suggested algorithm would flag. And I’m not sure something like a “hot” label would be a good idea in any case. If knowledgeable IDers think a particular observation is “difficult”, I would suggest we almost want to draw attention away from it. Experts will eventually clue in to how/where to find this type of observation. (In general, we know each other and we’ll let each other know.)
Hiding difficult obs under the rug until resolute IDers lift it? Well… A basic principle of the iNaturalist platform is that everybody is equally entitled to “be bold” (1st item of iNat ID Etiquette) and to add identifications as they please, neither expertise nor justification required. Keeping control of ‘difficult’ observations is already slightly frowned upon; I wouldn’t hold my breath for additional features to this avail. Which is why I suggested a possible UI trick more in line with the “accept community input” law of the land, a trick that would automatically invite extra/extended scrutiny on markedly difficult obs.
In a scenario like this, I think it’s reasonable to apply iNat’s existing tool for this case, the “No, It’s as good as can be” field. If multiple experts have concluded that it can’t currently be IDed past genus, it is fine to tick this box (though I know that there are a variety of opinions on this). The observation then becomes RG (and on GBIF) at genus, increasing it’s value.
As noted
IDers can tag other experts or experts can search for observations that are RG at genus. This approach is more efficient than leaving these observations in Needs ID forever, reduces the risk of an overzealous IDer erroneously providing a more specific ID (as described above), and uses existing iNat tools. To my mind, adding additional tools for finer and finer distinctions of observation “grades” for more specialized use cases that increasingly smaller proportions of iNat users will likely make use of isn’t the most effective use of limited staff time/effort for developing the platform.
I guess one problem with this is that I’m usually one of the first to look at observations in my region/taxon of interest. To some, it may seem a bit premature if I set the ID to genus/complex level and click “as good as it can be” before anyone else has even had a look at it, even if I add an explanation for why.
Revisiting these observations at some later date to set the “as good as it can be” field is additional work, especially since we’re lacking in ways to manage multiple work queues at the present time. In addition, waiting for other knowledgable IDers to chime in with their thoughts runs the risk that in the meantime, someone less knowledgeable will add their IDs.
I guess maybe I should “be bold” and just set the “as good as it can be” field as I see fit. Even if there’s blowback, why should I let it bother me? It appears everyone is free to do pretty much as they please. I can use the metadata tool to do this, so that certainly speeds things up.
Those are your words, not mine. I said “almost want to draw attention away from it”. That said, doesn’t applying the “ID is as good as it can be” flag pretty much hide difficult observations under the rug?
My suggestion was to leave difficult observations as Needs_ID, but add some kind of a flag indicating that it’s in that state because it’s difficult to ID. Having an alternate status like “Needs_ID (is difficult)" would move these observations out of the “Needs_ID” queue, so it would be less likely that they would get casually ID’d by users who are trying to cut down the Needs_ID queue.
Note that I started off by saying that I didn’t like the emphasis that is placed on reducing the Needs_ID queue. I personally don’t think it’s a problem if some observations remain in the Needs ID state forever. Encouraging folks to cut down the “needs ID” queue just trades one problem for another. Instead of observations languishing in the needs_ID state, you trade them for observations in the RG state that may or may not have the correct ID (because those who pushed them to RG did so for the sake of getting them out of the needs_id state, regardless of whether or not they had the expertise to correctly ID the tricky ones). I review/correct/annotate observations in my region/taxon that are in the RG state, but I suspect that many don’t. After all, cutting down the needs_ID backlog is the priority, right?
I see that observation had been entered in the Phylogenetic Project for Unknown/Magnoliopsida, but that doesn’t narrow the search much. Conceivably, it could have been a much easier search if it had been sorted into the project for Tribe Calenduleae, but the odds were stacked against it.
- This observation has no flowers, so the only feature that the CV could go on was the leaves.
- There are 66 species of Osteospermum with observations on iNat, but only 20 of those species have enough observations to be included in the CV model.
- O. pyrifolium has leaves that are distinctly different from those other species, so there’s nothing the CV can use to say this observation is similar to them. I presume that’s true when comparing it with the rest of the tribe.
You and those other identifiers you tagged may well be the only experts on iNaturalist who could have identified this observation. If none of you noticed it, I don’t think any system tracking reviews by non-expert identifiers would have helped get it to your attention.
If someone knows a way to double the number of expert identifiers on iNat, for every group of rare species on the planet, I’m sure we’d all love to hear it.
In the meantime, congratulations on rescuing this rare observation from obscurity!
My experience is that with taxa that people want to ID to species even though it isn’t always possible from photos, they will often add a species ID regardless of whether or not someone has clicked the DQA to make it RG at genus. So while using the DQA helps act as a deterrent, there are a meaningful number of users who do not pay attention to RG vs. Needs ID when adding IDs, for any number of reasons.
I am unconvinced that people adding unknowledgeable IDs to difficult observations are doing so primarily in order to reduce the Needs ID pile. The people who seem to be most concerned about this are people who are systematically reviewing a taxon and thus generally people with some expertise.
(Note: I spend a lot of time IDing taxa where a large percentage of species can’t reliably be identified based on photos and thus many many observations stay in Needs ID indefinitely. This may be different for taxa where a species-level ID is feasible for most observations and thus Needs ID vs. RG is useful for managing workflows.)
But - my point is - if iNat guidelines say - dump it in a broad ID - then iNat needs to motivate someone else to LOOK at it again. It is a different skill set to a taxon specialist combing thru - and would be an appalling waste of their time. Unless like me they enjoy mudlarking.
I think it’s going to be taxon dependent, as you suggest. I’m doing butterflies, which is a familiar/popular taxon that’s likely to attract a lot of casual identifiers. I don’t see notifications of people agreeing with existing IDs, so there’s a big chunk of missing data there. I think I do get notifications for IDs that are added to observations that I have only commented on or annotated. I do sometimes get a few notifications of people who are adding IDs to observations that went RG long ago. Anecdotally, these are mostly just agreements (which I find mystifying). It’s pretty rare to see someone going through old RG observations and disagreeing with the existing IDs. I guess maybe that means we’ve been doing a good job with the IDs, but there has to be something there that someone will want to disagree with. There has been some of that very recently, where somebody with some expertise in a particular genus has been combing through both the NID and RG observations and cleaning up some of the ambiguous cases. I noticed this precisely because I don’t see this happening very often. Of course, because I haven’t been using the “as good as it gets” option, we don’t have very many observations that are RG above the species level, so I don’t know what the likelihood is of non-experts trying to ID these to a lower level. I have little evidence either way. But I guess I’m going to find out.
A lot of the “interference” I see stems from random folks who add IDs before I get a chance to look at observations. Pretty much every morning, I check for new observations that have been added to our project overnight. But still, there are random people who beat me to the punch, and enter an incorrect/too specific ID. Frequently, they are identifiers I’m not familiar with, who are from far outside our area. They never participate in any discussions around their incorrect IDs, and I typically don’t ever see them again. This is why I refer to them as “drive-by” IDers. Another bit of evidence is that they tend to only contribute IDs on fairly recent NID observations. I don’t tend to see notifications for people adding lower level IDs on very old NID observations (of which there are plenty in my region/taxon).
There is also a small contingent of folks who I do recognize. They are from far outside our region, so they don’t have good understanding of some of the intricacies of problem taxa in our region. They contribute a few IDs, and then they seem to disappear for a several months. My hunch is that they “drop by” whenever they have time on their hands. I’m sure there are plenty of NID observations they could work on that are in their own region - likely far more than there is in mine (because I keep on top of the observations in our region). My guess is that their search is simply based on recency - they have some time, so they look at the NID backlog for the continent, and start working backwards. I don’t see these folks going after observations that are already RG (though as I said, we don’t have a lot of higher level IDs that are RG).
There is a small cadre of folks from outside our region who DO have expertise, and regularly contribute IDs. They tend to be focused on particular families or subfamilies. These folks are genuinely helpful, and their contributions make my life easier. I don’t want to give the impression that I’m ungrateful for the contributions of folks from outside my region.
Because the people whose contributions cause me grief tend to be semi-random, and they mostly contribute IDs on recent NID observations, I’m led to suspect that these are well-meaning folks who are looking for NID observations they can get out of the queue. Given the emphasis on keeping the NID queue under control, this makes sense. Frequently, you get what you ask for.
providing identifiers with better tools for managing their workflow might help. I stopped doing ID work on a competing platform because of the overall poor quality of the software. Conceptually, the platform is better than iNat in some respects, but the execution is horrific. Basic functionality simply doesn’t work, or works erratically. Because what little effort they put into testing and bug fixing is directed to the submission side of things, it is particularly bad for anyone trying to do ID work. I used to be a major booster for the platform, but they succeeded in converting me into their worst critic. The iNaturalist platform is infinitely better. In general, the basic functionality works as advertised and one can get things done without a lot of hair pulling and cursing. That doesn’t mean it can’t be improved.
I suspect a lot of experts will take one look at iNat and conclude that it’s more “pretend science” than “citizen science”.
Be careful what you wish for. It may come true.
Well, but, citizenship is relevant neither to observing nor IDing. We had an earlier discussion around that fraught topic.
Yes, if our overworked specialists can’t spot a new observation quickly, it gets buried in the pile of unknowns. If someone follows the ID-a-thon suggestions and sorts that observation into the only slightly less daunting dicot or monocot piles, it can get buried under those. And I don’t know if there are enough intermediate level identifiers who can sort those a step or two further, to order or family.
I looked at your silver spoon example, again, and I can’t even notice any veins in the leaves that would have let me identify that as a dicot. (I only see one central vein, or is that just a central fold?) Is there any guide that would possibly have led a non-expert to get that down to family?
Given that it is a local Endemic, my own preference would have been to leave it Unknown. Those do get looked at.
Putting it in a clean box, with a new label is effectively archiving the obs, without adding much to the ID. It is a daisy, weird and unusual ( = local Endemic). For daisies in Eden there are 722 sp to choose from - which a taxon specialist would / could consider from Family, but they are not going to pick over all the flowers on the off-chance of finding a daisy. You would find some, but … and 26K NID daisies waiting in Eden already.
In the 1700s, “mudlarks” were poor children who scavenged the mud of the River Thames, at low tide, in order to find bits of coal, metal, or other scraps that they could sell for a few pence a day. It was a dirty job! Just like trying to ID certain types of observations on iNaturalist.
“Mudlark” was a mocking term, which was possibly a riff on “skylark” (a bird known for singing “joyfully” while “larking” in the sky). So “mudlark” referred to the idea that, instead of “larking” in the sky, these poor kids were “larking” (grubbing around) in the mud.
Sewers and plumbing would not appear for 150 years, so they were wading in human feces. Hopefully, observations that are poorly photographed, and confusing, and hard to ID, don’t elicit the same revulsion, but they do “spark joy” when you get them to Research Grade.
Revising erroneous RG obs based on textures and questionable hosts.. sometimes having to use visual adjustments on photos to tease out the details. I get it. The dredge is equally thrilling.
Often comes up in novels set in old London. This book fascinated me.
Elif SHAFAK - There are rivers in the sky
You know, forum users are capable of looking up the definition of words they are unfamiliar with themselves, or they can ask the poster what meaning was intended. They don’t need an off-topic commentary accompanying every post in which someone uses a word that may not be widely known to everyone.
I could understand your contention but I enjoy footnotes. I only see an issue if theres sustained dialog that stays from the original topic.
( @dgwdoesthings Adding here to avoid another off-topic comment about off-topic comments: Not invalidating that position necessarily, as long as the information is correct and in good faith without significant derailment it may be pedantic to be any more than slightly annoyed. I digress.)
At the danger of going off topic… The issue that spiphany is higlighting is a repeated behavior where a sidenote is provided without a user requesting explanation and often the responder does not know before hand and contains text sourced from generative AI and not from the users own knowledge.
I’ve also found it mildly annoying and distracting but not enough to say something about it, plus it’s not really against any forum guidelines.