Increasingly it seems like more and more of these are getting imported by users using the external imports tools. When they show up in the ungrafted taxa list, I am curious what approach other curators feel should be taken.
Guidance on these less official node levels in the curator guide can I think be best called as cautious, leaning towards preferring against their use :
if you graft them into place, they just end up as orphans, nothing moves under them. It can be a huge task to populate them, even worse if someone imports subfamility X, but it turns out subfamilies Y,Z, A, B, C and D also exist
set them as inactive, which will inevitably lead to flags about why did you set my subfamily inactive, it is a perfectly valid name.
I generally take the first approachāgraft them in place, and if I have time and understanding to work it out, try to move some or all of the correct genera. The major failure cases here are people following the taxon, or using it in identification filters, and not capturing some observations due to genera not being moved; or identification disagreements getting bumped up to a higher level. But people making use of subfamilies and tribes are likely to have a good grasp of the taxonomy on that branch of the tree, and this will presumably help bring it to their eyes and encourage them to fix it, so I think itās preferable to suppressing those ranks.
That on is kind of the biggest hurdle in my view. The curator guide is pretty clear that we should adopt an all or none approach, ie donāt start what you canāt finish.
Specifically it says āIf you want to add additional nodes to a tree make sure you are willing and able to spend the time ensuring that all nodes globally that should be included in that node are grafted as descendants.ā
And further āAs a rule of thumb, not including additional nodes in iNaturalist is preferable to including but only partially curating additional nodes as in the middle tree above.ā
I had taken that as a directive on curators importing new taxaādonāt deliberately import intermediate levels of classification unless youāre willing to populate them. OTOH, once theyāve already been pulled into the database, I think a partial fix is better than none.
Glad I found this conversation! Itās almost 2 years old, but it still seems important to resolve. Itās what I headed to the forum today to look for - @kitty12 informed me, on a flag on a name I added, about the relevant iNat guidelines and I was surprised/disturbed to see their current state. This is a more complex problem than I thought at first, and it has taken me a while to figure out how to explain my intuition. I donāt think we canāt resolve whatās best here with any one āhereās a bad outcome that can happen with the ruleā or āhereās a bad outcome that can happen without the ruleā.
The guidelines as currently written are clear about how they want us to handle this situation:
In other words: for extra-KPCOFGS taxa, when the options are removing/not adding the taxon and having the taxon, but without all existing iNat subtaxa being grafted on, a curator is to choose the former. Why? The explanation in the Guide is (pretty much only) this, which Iāll call Scenario 1:
The tl;dr of my reaction is this: Prohibiting (any) useful names for (any) period means permanently losing out on valuable knowledge/information that IDs using those names would have provided. Most affected observations will be, permanently, that much less precise/accurate. Itās well worth the price of having unnecessarily imprecise (not inaccurate) observation titles for that period. As the names are eventually completed, observation titles are automatically made (perfectly) precise and (perfectly) accurate by iNat. An identifier who was denied access to the name before probably isnāt going to come back to re-identify this observation when the name is finally added.
The too-long version of my reaction is this:
We have to acknowledge that eventually, all of these names (that meet other criteria for inclusion) WILL be added to iNat, with all their subtaxa grafted. If itās a valid, useful taxon, eventually someone or some people will have the time and effort to add it and graft on all of its subtaxa.
So, letās compare āillegalā (the current rules; incompletely subtaxified extra-KPCOFGS names are illegal; the curator wouldnāt add the name at all) and ālegalā (the proposed rules; incompletely subtaxified extra-KPCOFGS names are legal; the curator would add the name but leave it incompletely subtaxified).
Letās look at Scenario 1 from above, which is supposed to be likely/bad enough that we canāt allow it to happen. In reality, genus Garnieria is a member of tribe Persooniae, which is a member of family Proteaceae. Genus Garnieria and family Proteaceae are already on iNat, and a curator whoās short on time is deciding whether to add tribe Persooniae, under family Proteaceae (but neglect to graft on genus Garnieria, for the time being), or not to add tribe Persooniae at all.
On one particular observation, user A would like to propose tribe Persooniae; user B would like to propose genus Garnieria. Would Aās tribe proposal be a confirming or disagreeing ID with Bās genus proposal? A proposes first, so wonāt be clarifying it to iNat, but we can consider both intentions and look at the cons resulting from the current and proposed rules.
Under āillegalā (the current rules):
Scenario 1a. Aās tribe Persooniae would be meant to confirm Bās genus Garnieria.
A canāt make the tribe ID (that name is illegal). So, A settles for proposing family Proteaceae; B proposes genus Garnieria. Title ā genus Garnieria.
Immediate effects: A and B are both happy with the title
Long-term effects: Aās belief that itās tribe Persooniae (not just family Proteaceae) is permanently lost, making the outcome of future proposals that much less accurate/precise/confident.
Scenario 1b. Aās tribe Persooniae would be meant to disagree with Bās genus Garnieria.
A canāt make the tribe ID (that name is illegal). So, A settles for proposing family Proteaceae; B proposes genus Garnieria. Title ā genus Garnieria.
Immediate effects: A is unhappy with the title (but this is just because A proposed first & doesnāt get to specify disagreement, not because of the rule); B is happy.
Long-term effects: Aās belief that itās tribe Persooniae (not just family Proteaceae) is permanently lost, making the outcome of future proposals that much less accurate/precise/confident.
Under ālegalā (the proposed rules):
Scenario 1a. Aās tribe Persooniae would be meant to confirm Bās genus Garnieria.
A proposes tribe Persooniae; B proposes genus Garnieria. Title ā family Proteaceae.
Immediate effects: A and B are both unhappy with the title for the time being - they both would have wanted the much more precise (but no more accurate) genus Garnieria.
Long-term effects: Once tribe Persooniae is properly subtaxified, the title will change to genus Garnieria, and both will be happy.
Scenario 1b. Aās tribe Persooniae would be meant to disagree with Bās genus Garnieria.
A proposes tribe Persooniae; B proposes genus Garnieria. Title ā family Proteaceae.
Immediate effects: A is ambivalent about the title for the time being (tribe Persooniae would have been better, but at least itās not genus Garnieria, which they believe is wrong). B is unhappy with the title (but would have been equally unhappy under āillegalā if A had gone second and been able to express their disagreement, so not unhappy because of the rules).
Long-term effects: Once Persooniae is properly subtaxified, the title will change to Garnieria, and A will be unhappy (but just because they IDād first & didnāt get to specify disagreement), and B will be happy.
It seems to me that in both scenarios, the ālegalā outcome is better. Yes, the observation will get a unnecessarily, significantly (Scenario 1a) or moderately (Scenario 1b) less precise title for the time being. But that title is on the basis of more, valuable information, which is now being preserved, and will be immediately, properly, automatically taken into account and reflected in the observation title, as soon as tribe Persooniae is fully subtaxified. Which we know will happen. Not only will it happen, but it will happen sooner if it can be done piecemeal than if it has to be done all at once. Any of us curators can find 2 half-hour periods to work on iNat taxonomy in a shorter span than we can find 1 hour-long period. Letās not let the perfect be the enemy of the good, and add the name now, and work on integrating it as time goes on.
By the way, we should keep in mind this scenario was cherry-picked to showcase the āworstā of the ālegalā protocol. A perhaps equally (more?) common scenario would be
Scenario 2. User A wants to propose tribe Persooniae. There is no user B who looks at this observation. Under āillegalā, A canāt do it and has to settle for family Proteaceae. Under ālegalā, A can do it. Simple.
Iām really curious to hear othersā thoughts on the issue in these terms, because despite this wall of text, it really seems like a no-brainer to me. Collect as much valuable information as possible, and donāt reject it just because it makes the interim display less pretty (less precise, not less accurate). I come to this from the context of Mushroom Observer users importing our observations en masse and adding the extra-KPCOFGS names weāve used there, here, not wanting to throw away the precious information contained in those IDs. But I think it applies equally to any curator adding names.