The “good as it can be” box is incredibly useful in cases where the OP made an incorrect original ID and then a bunch of inexperienced users piled on to agree with it (as happens a lot with “duress” student project observations). If an identifier adds the correct ID with, say, six incorrect IDs to overturn, it will stay at RG…unless that identifier ticks the box, which greatly increases the chances that other identifiers will see the observation and add the correct IDs necessary to overturn the incorrect IDs.
The box is also very useful for photos in which details vital for ID are not visible, so even a specialist in that taxon can’t ID to species. It enables RG observations at genus level, which is very useful for some groups.
And (as others have touched on) it’s a way of “frassing” photos that are so lousy that it’s not even clear what the subject was meant to be.
I sometimes use it. If the photo is blurred, or anyway diagnostic characters are not visible, there is nothing that can be done past a certain level. I consider it a service to my future colleagues so they don’t waste time.
I don’t really understand what to do with the observations in this project.
I looked at the good-as-it-can-be observations from the coast of Ecuador. There aren’t many, and only one that I could id. It is the only observation ever posted by that person and consists of three photos, each of a different species. I can put a name to all three, and did so in the comments. I ticked the yes it can be improved box But what else can be done to fix it? The observer is long gone.
That is one for the new DQA (at the bottom of the list) - Not a Single Subject.
Which is automatically Casual, and out of the Needs ID queue.
If the observer ever returns they can resolve the problem.
That is not a a case for yes/no Can Be Improved (but the current Yes and No cancel each other out, which is good)
Ok, that is a good solution in this case. None of the three species is anything to write home about. But what if one had been something noteworthy? What if it was a Wow, that’s a first!
There’s a ton of turtle photos that can’t really be ID’d beyond subfamily too. And a lot of Plestidon photos that can’t get further than genus due to similarity of juveniles where species overlap. I do think that the button serves a real purpose and there’s a lot of observations where it’s valid at a taxonomic level other than species
As an identifier, I really disagree with clicking ‘good as can be’ except for the most blurry photos. I can say that I can’t identify it today but to say that no one else could identify now and now one ever will is too great a leap. Identifiers learn more all the time. Half the species I can identify today I couldn’t identify five years ago. Should I have marked them all as ‘good as can be’ at that time? Having to see them again all the time is a benefit not a draw back. And how could I be so bold as to assume that no one more knowledgeable than me could potentially be out there? Someone might see it differently today, or someone else might join the site tomorrow. I’m glad that the species I’m interested in weren’t all marked as good as can be before I joined the site.
I wholeheartedly agree!
There are many mushroom obs held back by this. I’m trying to go through them but without a way to filter out DQA markings they’re found by chance.* Some just need a “yes” tick to hit RG, others are in the Casual purgatory.
*if there’s a way to find this in the API search page, I’m interested… otherwise, I’m not tech-y enough for other means.
In the OP, Diana mentioned the phrasing of the DQA. I agree it could be improved, but we have to separate the phrasing of the label – the necessarily short word or phrase that has to fit on the DQA form – with the phrasing of the longer text that explains the label, which would have to come from a Help page or (more accessibly) from a pop-up help window that might appear if we hover our mouse over that label. And if we’re going to fully explain all of the situations and edge cases that have already been mentioned in this thread, that help text might have to be much, much longer!
I’d like to see this DQA used more often to help reduce our daunting pile of Need IDs, so this is more like how I’d interpret it. The expertise required to make a good-faith, reasonable, and provisional declaration like this would depend upon the observation and the current level of the CID. If the ID is for a genus with very subtle differences between species, then we might have to leave this decision to the experts. On the other hand, if the photo is a landscape of distant trees, where the observer hasn’t even offered their own ID to indicate a particular object for the observation, how much expertise would we need to declare that an ID of Plantae can’t be improved? Who would qualify as a “distant tree” expert, anyway?
In either case, if someone does come forward to attempt these IDs, the important thing would be that it’s easy to find the set of observations with this DQA checked. It’s great that @jeanphilippeb has made a project to collect these observation, although my (vague) understanding of this is that the API calls to manage this project and all of the phylogenetic projects are a limited resource. It would be better if iNat would provide a filter for us to be able to search for this DQA through the Explore or Identify pages.
If there’s a Feature Request for this that I haven’t found, I’ll vote for it. If there’s no outstanding Feature Request, I’ll try to compose my own.
Female Axarus festivus complex members are not able to be IDed without dissection. The species are identified by male genitalia. This is an example of something that should be marked as good as can be.
I’ve gone through and used it for two treefrog species that overlap in my area and you can only differentiate by call. If there ever is research that shows a new way to tell the two apart, then they will be ready and waiting at the species complex- but they will have been useful as research grade in the meantime.
We can say that even the blurriest photos have the potential to be IDed or might be valuable to someone, but the truth is that there are many low quality observations that have a miniscule chance of ever being IDed and aren’t super useful even if they get to RG (really commonly observed species with clear range maps come to mind.) At some point, you need to weigh cost benefit of a possible future identifier getting it to RG vs clogging needs id for years. Maaaybe somebody else will be able to magically figure out what type of pine that blurry green blob is, at which point it will join ten other observations of the same three trees all taken for a school project.
I’m still fairly conservative about this use, but I think sometimes it’s better to just make the call.
I click “No, it’s as good as it can be” for all weedy dandelions (Taraxacum) because the genus is a taxonomic mess. Before we consider whether we could ID the species, we’d have to agree on the species and people don’t.
Otherwise, I probably don’t use it as often as might be appropriate. A lot of distant grass and sedge photos probably can’t be ID’d, but I keep thinking “Maybe someone with very local knowledge might know this.” Distant dicot trees – many of them are simply not identifiable. I mostly just skim past, but calling them “good as it can be” would be OK with me. Also, I agree that most of the really bad photos are common species already well represented on iNaturalist, so making them disappear is not necessarily a loss. (Thinking class projects and CNC, mostly.)
I would only withdraw my ID if I felt that it was likely wrong, and I didn’t have a better suggestion to replace it with (even at a coarser level).
Withdrawing would simply leave the first ill-considered “agreeing” ID as the only ID, and one more ill-considered agreement would send it to research grade.
By voting Yes, it can still be confirmed or improved, I keep it in the Needs ID pool until it attracts enough IDs that I trust (agreeing or otherwise). Then I would remove my vote, even if there wasn’t yet consensus on the ID.
Of course, if someone decides to counteract my Yes vote, then other options come into play.
Don’t get me wrong, I don’t use this option a lot, just in cases where I want more eyes on the ID before it becomes available to GBIF.
Thanks for pointing that out. For posts in the Feature Request category, we can vote for requests to show our support. Will liking that post (one of many suggestions in a long thread) be recognized in the same way?