Thank you so much for all your summary posts!
And wow, everyone accomplished quite a lot in a month - NO older Unknowns left? That’s astonishing.
4 Likes
I put the raw data in a comment in the journal post if anyone wants to have a look themselves https://www.inaturalist.org/posts/83883-final-numbers-update-2023-09-01#comments
2 Likes
Back at the beginning of August I was pegging away.
Thanks to August identifiers. Who have become September identifiers.
Today my African Unknowns are zero! Never thought I’d see the day.
8 Likes
Thanks for the nudge about those Araceae.
1 Like
This is the first half of what I was waiting for. The last 10 August days will be in the next / October update.
https://www.inaturalist.org/posts/84723-a-new-computer-vision-model-v2-7-including-1-785-new-taxa
trained on data exported last month on August 20th
The URL is clunky - but if you start from the internal Plantae link. Then filter for Africa. iNat has added a perfect one hundred plant sp for Africa!
7 Likes
Wow, that’s great!
1 Like
Deliberately adding an incorrect ID to an observation is a violation of iNat terms of service
If it is not Life.
Then it is No Evidence of Life. And that is wrong in my book.
But I will no longer mark them ‘Good as can be’ as that is generating broad planty IDs at Casual. Which serves no useful purpose.
Marking a photo of a plant as no evidence of life is a violation of the iNat terms of service.
I was looking for a link, and saw these instructions that violated TOS, and was concerned someone might come along later and read them without context
My biggest issue with “Life” is that it essentially bumps things all the way back to unknown. State of matter Life is included by default when you select unknowns for identifications. There is a URL hack to exclude it, but how many people know about it? If the goal is to reduce the number of unknowns, whether Needs ID or Casual, it seems counter-productive to that goal.
2 Likes
There is also data loss in Lifing some duplicates vs others. Examples:
- Hand-added annotations (valuable metadata) are obliterated, and they do not return if the proper taxon is restored afterward. This matters if someone annotated one dupe but not the other.
- A particular issue is if “duplicate” has been only loosely interpreted by the Lifer, and the data loss is of things like revealing additional camera angles.
- An unintended consequence is when the observer deletes the other obs vs the obs someone ordered them to delete. Meanwhile the other remaining copy is mislabeled in the Life pile, sometimes helpfully labeled casual too.
- An unintended consequence of Lifing strife is that some folks up and leave such strife. The ultimate important impact of this fun challenge was that a valuable teammate performed a self immolation in protest of injustice.
Due to the last bullet point, I feel the mission status should be: FAILED
At least until we fix things. One way to fix? Agree to best practices, and write them all down where someone can read them and keep linking to them. Like in the Etiquette thread in OP maybe?
Meanwhile, I’ve seen the raw research data coming right out of the bush. It ain’t always pretty. Maybe it’s hard to push the right mobile buttons in the glare or heat or insects whatever. My strong advice us to try to push all dupes (however interpreted) along at the same level, for ease of analysis later.
When is later?
- When I get back to it. ;)
- When the observer gets back to it.
- When the AI-powered identifying interface that we’ll command will make meaningless the distinction of how the observer has binned their photosets. Making our present disputes moot.
5 Likes
There are staff statements about preferred practices for multiple-species observations here and about duplicates here. The guidance on the first has been added to the FAQ page; @tiwane indicated he was going to add one about duplicates, but I don’t see it on the page yet.
I agree that additionally putting these things on a page more oriented towards the needs of IDers rather than buried among general “about iNat” questions would probably help make them more prominent and findable.
3 Likes
I would like iNat to label Placeholder as Temporary
But my request was turned down, since iNat staff know it is temporary.
Since the observer added the placeholder text - and sometimes it is a chunk of info - the observer should control if and when their text is vanished.
Whatever ID you add to move it from Unknown … the placeholder text is a booby trap which vanishes.
If you are in the field with no internet access, it may be vital to label ‘observation 3 as Possibly whatever’ If it is a painfully sourced Missing on iNat sp we need to keep the info visible on iNat. Somehow.
1 Like
If we want to make iNaturalist more focused on the data needs of researchers, then we will have to completely revamp the way it is promoted to the public. Of course, that would also “solve the problem” of the tsunami of observations, if you get my drift. When the user base realizes that they are expected to be a workforce of unpaid field assistants, how many will stay on?
1 Like
If that were really the case, then yes, it would certainly be pretty “selective”. But I do think more emphasis could be put on the idea of community, with all the give and take that involves. Far too many users seem to equate iNaturalist with a simple one-way app to get a quick ID and are unaware of - and often uninterested in - all the work that goes on behind the scenes. Reinforcing that concept would not be at all a bad idea.
4 Likes
I am sad that @lotteryd has failed the whole class.
Thank you to the identifiers who kindly invested time and effort on African plants thru August. And especially to @peakaytea who remains a daily companion among our African plants.
2 Likes
The particular conflict that happened here had nothing to do with the data needs of researchers vs. the experience of lay users.
It had to do with inadequacies in iNat’s design that don’t provide a good mechanism for dealing with observations with problems (something that is in everyone’s best interest to be able to resolve – including observers, who would probably have liked an ID for the observations they innocently uploaded incorrectly).
And it had to do with certain users gatekeeping and aggressively enforcing unwritten norms which reflected those users’ personal preferences for dealing with certain situations (not objective research needs). What matters is that this bullying was allowed to happen unchecked and the person experiencing it was left feeling that there was nowhere to turn and no way to mediate the conflict.
Conflicts are inevitable in any community. What is decisive for whether they survive or dissolve into ever-increasing strife is how they manage that conflict. iNat failed in this case.
3 Likes
I’m sorry, I think I have lost track of the thread of conversation here - what do you mean by saying lotteryd has failed the whole class?
That makes me three times sad. For the one who left. For lottery and her shiny project. And for the hard working identifiers.
You’ve taken it out of context, and completely misquoted. To what purpose?
Anyone interested should follow the link to read the post.
Many of us can work on the things lotteryd talked about.
1 Like