Identifiers as Volunteer Curators, and Obs with Probs

[tiwane]

First, I think we should remember that everyone here is trying to do the best they can to improve iNat, even though we have different philosophies and ways of going about it. I think it’s also best to acknowledge that there will always be messiness, gray areas, and edge cases.

@spiphany

multiple unrelated photos showing completely different species (for example, a rose, a giraffe, and a pigeon) in one observation. This is a relatively common mistake with new users who don’t realize that they need to create a new observation for each organism.

With the upcoming mobile app, I suspect this problem will be reduced due to the new observation workflow and guidance in it. It will still happen of course, but I think the occurence will be greatly reduced.

@spiphany

I would go further and suggest that an informal consensus is not sufficient: there needs to be some official guidance on how IDers should handle observations that have issues for which there is no obvious, standard solution such as flags or DQA.

There’s official guidance here, although it’s buried and should have its own FAQ.

If you see an observation that has two or more photos depicting different organisms in each, it’s best to identify to the level that fits all photos and make a comment politely asking the user to separate the photos into different observations. For example, if the first photo is a flower and the second is a bug (without the flower), identify it as “Life”. If the user is unresponsive to requests to separate the photos, mark the observation as “Based on the evidence, can the Community ID still be confirmed or improved? No, it’s as good as it can be” in the Data Quality Assessment section.

I can make a new FAQ for that today.

@cabintom

Edit: I guess I just don’t see any practical upside/benefit for IDing to the lowest commonly shared taxon.

I totally understand where you’re coming from, but I’d say there are a few issues it causes:

• observations record encounters with individual organisms, and the photos/sounds attached to the observation are evidence of that encounter. So IDing only the first photo is going against the system and definitions everyone is being asked to follow, and on which iNat is based, and it’s an inaccurate evaluation of the evidence provided.

• Because observations are defined as such, all photos from an observation are included in the taxon browser for the observation’s taxon, which creates confusing and weird situations. The photos might be used to train the computer vison model as well.