Hey all, as a follow-up to our ongoing work around surfacing identification expertise on iNaturalist, today we’re releasing the first version of a new tool called Identification Tips.
You’ll now see an Identification Tips tab on species-level taxon pages where community members can nominate helpful identification remarks, vote on what’s most useful, and browse the results. Everything is linked back to where they came from on their original observations.
This is the first version of the tool, and we expect it to keep evolving as the community uses it and shares feedback. Please feel free to explore it for yourself and share your constructive thoughts via this form or in this thread!
Why do identification tips have to be attached to an identification, not a comment?
Building on the existing structure of identification remarks means that every identification tip inherits useful context automatically: who made the identification, what species they suggested, and a link back to the original observation. Some of these things can be traced back through comments, too, but we wanted to make sure every tip is clearly linked back to an independent identification.
We would need to add a field to the comment for the ID – to someone reading IDs and comments on an obs it is obvious that – John IDed that as Some species – belongs with John’s following comment.
It is effectively carving out a substantial chunk of useful info if we cannot nominate comments. I am tempted to copypasta ‘John say this sp because’ but I won’t because then it will be tied to Diana says. Which is not as useful.
I understand why the current limitation (ID comments, and not lone comments) exists. I do think it’s better to test-run this concept on a smaller and more regular test pool, any way. It might be possible to later add comments into the mix, but why make the initial trial more complicated than it needs to be?
Observer submits record and adds identification as species 1
Identifier adds species 2
Observer ask “why not species 1?”
Identifer adds comment “because …”
I only provide clarification when asked. I don’t have time for anything else, and I’m sure that is true of most prolific identifiers. Consequently, the useful information will nearly always be in comments.
interesting. if i were going to poke at this, i would try to see what happens when taxon changes happen and when people delete observations / identifications / accounts. maybe i’ll look more when i have more time…
first impressions, i guess this way of doing things helps to avoid some of concerns that people had about previous attempts to get this information, while still incentivizing people to make good identification notes. so i think think that’s probably the best possible balance of those things. i wish each note would have a little more geographic and temporal context in the identification tips tab.
in terms of the voting mechanism, i guess this is more or less what i asked for in terms of having something similar to how StackOverflow works. i think there are some inherent issues with the approach, such as difficulty in getting newer comments voted to the top, but i’m not sure there’s a better way to do it at the end of the day.
i like that there’s an API endpoint /v2/exemplar_identifications, too.
it looks to me like every ID comment ever, but i could be wrong. once folks start nominating identification notes, it’s easier to sort through just the nominated notes.
i would do both, if you think it’s something that definitely needs to get in the final list. but i can also see a case for just nominating as a means of just collecting even remotely relevant notes as a first step in the process.
maybe they could add separate nominate + upvote and nominate actions?
one situation that i think i might use just a nominate without an upvote is to highlight unusual forms and cases. these wouldn’t be the best notes for usual identification, but they still might be useful for identifying unusual cases.
Yes, right now, all remarks ever made are being presented! There’s no initial filtering on this release, and manual input/community voting decides what gets surfaced.
In the future we’re interested to explore ways that we could potentially use AI/predictive models of some kind to surface the most useful comments (i.e. through some kind of helpful sorting mechanism that can then be refined by community input), but those explorations are farther down the road. One example that comes to my mind is that it would be helpful to have an easy way to filter out all the “welcome to iNaturalist!” comments that might show up; but again, the team is still in early explorations for the best way to make that happen.
Slightly regretting my habit of putting comments separately to identifications (especially when pinging people on my own observations to show a species / being asked after the initial ID), but a very nice feature!
I’ll be interested to see how this works with very widespread species or very variable ones - what’s an ID tip in one place (“this is the only species with x feature…”) often doesn’t apply elsewhere.
You can do what feels most comfortable for you, but keep in mind that logged-out folks won’t be able to see tips unless they’re upvoted! Let me know if that makes sense
Now I see it, but here’s the problem: you’ve made it so that we can only nominate IDs that have comments attached to them. The pattern I often see is this:
Person A posts something.
Person B identifies it.
Person A asks a question about the ID.
Person B explains it — but in a comment, not an ID.
In this scenario, apparently nothing is eligible for nomination, even if step 4 was the best explanation ever written about step 2. [EDIT: I now see that @cooperj outlined the exact same situation a few minutes ago.]
(By the way, simply “Nominate” under the little menu is not at all obvious. It ought to say “Nominate as a helpful comment” or something.)