if an observation with community ID turned off gets enough IDs of a different species so the original ID ends up as ‘maverick’, then it won’t display on range maps
2 Likes
This probationary period does eliminate the problem of inactive and short-term users (possibly students too) and I think it’s a great idea if we can find a way to incorporate it in a way that doesn’t discourage new users. However, I’m not sure that this solution nor better curatorial practices quite address the problem of what happens if a really active user goes inactive. I guess that side of the question is similar to the question of how to deal with IDs of deceased members. I’m just thinking, if I were to go inactive tomorrow, how would people deal with my identification as time goes on (just so you know, I’m not thinking of going inactive). Taxonomy changes, concepts of what populations are included in a given species changes. This is something we may not have to deal with much outside of the deceased category, but if a member moves on, are we going to need to adjust the way we deal with IDs?
Coming from the herbarium side of things, I tend to view these as citable records that should be flexible to changing times. In herbaria, the time element is taken into account by using the latest annotation. Since few people annotate and those that do are generally either experts in that group or curators the annotation quality level is already extremely high, so the latest almost always represents the most accurate name for the specimen. This isn’t really the case on iNaturalist, but we will still have to deal with the time factor if this ends up lasting for long periods of time.
3 Likes
Thanks–I didn’t know that.
2 Likes
Perhaps curators could have the ability to downvote problematic identifications of inactive users (inactive as defined as not active for at least 6 months, maybe longer)? If the user becomes active again, a notification could be sent giving them the option to reactivate their identifications. Or, perhaps the identification should be immediately be reactivated.
Outside of that, I don’t see much way to work around the scenario I mentioned above.
2 Likes
Well, assuming that iNat continues in some form or another, every one of us will go inactive for one reason or another - life happens, life doesn’t happen, etc. That would mean that in time, every one of the 17M (!) records currently on the site will lose their “owners.” This creates the potential of having a “standing wave” of valid accessible observations instead of one that constantly grows and accumulates over time.
This is one of the reasons I think that we should strongly consider changing the policy to be that, when someone posts an observation to iNat, they are donating it to the public good. This would be part of the social contract of participating in this community. For as long as they are current, active, contactable contributors they retain control of their observations to the degree that they can change them, turn off community ID (if they really want to), even delete them individually, but once they go awol, then control passes to the iNat community. Similarly, participants can choose to delete individual IDs and comments they made, but not do so for all their contributions en masse. If/when they choose to leave, the IDs could be anonymized, but not deleted. If this is not acceptable to a potential contributor, they have the option of choosing not to participate.
Nothing new here, but (I think) worth reiterating occasionally when these topics resurface :-)
10 Likes
I’d have to check the documentation to be sure my memory is correct, but I think if you do a split like that and the system can’t figure out based on range whether to place an ID in new A. bis or A. tris, it converts the old A. bis ID to the lowest common ancestor of both taxa, which in this case would be genus A. Which is basically what you want to happen, since if they’re sympatric, you’ll need to go through and look at each observation for distinguishing information.
The real problems seem to crop up when different observers are using the different circumscriptions of A. bis simultaneously; then it’s tempting to just try to manually trim it down to the new circumscription rather than do a split and worry about a bunch of observations suddenly jumping to genus level.
4 Likes
The taxonomy curation should adequate handle taxon splits. That seems to work very well already.
The issue of old identifications should not be an issue at all. That is how museums and herbaria work and have worked for a few hundred years.
There is a posting for flagging deceased users (which is bogged down on how to do this).
But dead and inactive users are not really the problem at all. The problem is not the users but the quality of their identifications. The only real way to deal with this issue to to use a reputation system. Beginners, once off-ers, and short-term users will not earn a reputation and therefore not unduly influence the identifications made by more serious users. That is also dealt with elsewhere on the forum.I dont see that any complicated system requiring curator input will work. We need a system that does it all in the background. We dont want a system where the word expert in Pogonophores comes onto iNat, identifies all 20 observations, and the system ignores him and his identifications because he has not posted ~20 observations, has not made ~100 identifications and was only busy on the site or ~2 hours instead of ~2 months.
5 Likes
Tony, if you want you could make a feature request for a reputation system (i don’t think anyone has done that yet?), though i think the devs are already thinking about it in the long term so maybe it will become a closed topic. I agree that the waiting period thing penalizes good users including those experts who don’t add observations just IDs, so I don’t think it’s a good solution either. A discussion thread in the general topic about reputation systems could work too. I won’t get too into that discussion here since it’s off topic.
2 Likes
There are problems with the reputation system model too… Just because you have identified 1000 observations correctly doesn’t mean the next one is going to be right. It’s kind of like saying flipping heads 10 times in a row means the next flip will be too… or must be a tails… you get my drift :) Nor does it mean that someone who identified only 10 things previously isn’t going to make a really sound identification on the next one. And how do you evaluate past experience and qualifications, it would be a nightmare!
@charlie do you mean the concept of probationary period when you refer to “waiting period”? If so, it would be less of an impact than waiting for a reputation system to settle out at your appropriate trust level!
3 Likes
yes i am talking about not supporting a waiting period. Not sure if i would have stayed on the site if that existed when i got here.
I’m not saying we need a reputation system either, whether i’d support it depends on how it works. Tony Rebelo wants one.
I can see a number of problems with copyright here, at least for observation photographs. As the law stands in most jurisdictions, copyright survives the person and remains part of their estate for a period of time (up to 70 years some places). The creative commons licensing system provides just that - a license, which can be revoked or modified at any time. The content remains the property of the individual.
1 Like
as far as i know the main issue is with observations being ‘locked’ out of community ID by absent users. I don’t think, rightfully so, that anyone is proposing we are able to manipulate other users’ photos.
3 Likes
Actually, this isn’t how museums and herbaria work (at least, no herbarium I’ve ever been to). The difference is the way in which the identification must be treated. In herbaria, the latest ID is used for filing. On iNaturalist, the most common ID is used (the community ID). It is a very different method with the later being less flexible across time, but much more robust to bad IDs. What I’m looking for is a way to make iNaturalist IDs more flexible with time.
The splitting taxa technique may work, but I need to play around with it more. I think most curators simply add an additional name and manually change the observations instead of using the split technique (I know that’s what I’ve been doing).
1 Like
Correct, at least on my part.
That all sounds like a good approach. In one of the examples given in the original post, @botaness made one observation in a couple of months, and two ID’s. 11 years ago. I don’t know if 6 months is too short, as some folks may be travelling & etc, but there has to be some cutoff time. A year inactive, without some explanation (I’m going to Antarctica for two years), could be used to override some controversial ID’s.
1 Like
The whole “create a new taxon with the same name when you change the circumscription” approach did seem pedantic to me at one point–but with the automated ID change powered by the atlases, it’s a whole different story. I have carried out a split as I described with Pleopeltis polypodioides, the resurrection fern, and it was much more effective than recruiting people to go around and manually re-ID as the new name. Highly recommended.
1 Like
There’s more documentation than most taxon splits for that example, if you want to use it as a case study for similar future splits, @nathantaylor. We really shouldn’t be doing manual IDs for most taxonomic changes, it should be done through the taxon change system whenever possible, because that means in most cases the inactive users’ IDs will be updated automatically as well.
https://www.inaturalist.org/taxon_changes/40444
https://www.inaturalist.org/posts/19073-splitting-pleopeltis-polypodioides-resurrection-ferns
2 Likes
At the risk of being pedantic, the herbarium uses the common scientific community ID (of which the latest publication is the current manifestation), whereas iNat uses general community consensus (without reputation).
Herbaria are thus up to date as of the last curation. But the power of community ID is that bad identifications can easily be detected, but that does not help if the community ID then cannot be updated because previous identifications provide too much inertia.
Flexibility with time is not an issue: the vast majority of taxa will stay the same, with some juggling of rank, and now with DNA a lot of juggling of clades. But the terminal units (species-subspecies) are relatively stable, although the naming may change (a trivial dictionary update). Splits which uprank IDs to the next taxonomic rank should solve these issues. We need to get the curators up to speed.
((In fact, I did it last night. A four way split of 72 observations: 64 observations remain as is, 1 taxon had no observations, 1 taxon had 1 observation, and the southern African taxon had 8 observations. So rather than split I just added the new taxa (and then belatedly discovered an Australian observation and another independently described taxon), and re-identified with reasons the relevant observations. I agree - a more efficient way is needed to deal with splits - a quick check: my effort changed all 8 to genus level community ID, of which two are now correctly ID’d - the others are still at the generic rank: using the inline split they would all be species-level already!))).
You are advocating a reputation system: one where everyone has the same reputation no matter what their experience, knowledge or understanding is.
The developers are playing with reputation systems. I suspect that the ones that they are envisaging (i.e. no external forcing, based on in-system use) are not delivering the results required to justify their implementation. But I may be wrong, and even if I am correct, a breakthrough may happen at any time, and would then speedily be implemented.
But I would be loath to start a discussion thread on reputation systems. Discussions on this forum tend imho to deal with exceptions, peripherals, problems and philosophy, whereas I prefer discussions that focus on solutions to well-defined, specific, solvable issues. Discussion of reputation systems would in my mind generate the same sort of debate as that about “TV” and “Smart phones” when they were first mooted, and now: how could we live without them?
i don’t really need you to paraphrase or pick apart my words here, and in fact, i wasn’t actually advocating either way between keeping the current system versus an implicit reputation system. That is beyond the point though. I asked you to make a thread because constantly nudging every irrelevant thread into a discussion about the reputation system you want is going to push things off topic, so if you want to discuss it, you can do so in a different thread, otherwise if you keep mentioning it elsewhere your posts may be moved or deleted when they are off topic.
1 Like