While I can’t answer your API-related questions, I hope I can provide some clarity as to fern taxonomy and what’s likely to happen in the future.
Ferns are indeed an “open” framework. I seem to be the de facto curator thereof by virtue of having put a lot of time into it, but I have no formal authority to speak for iNaturalist or make people do things my way. While the framework is linked to POWO, they adopted an arrangement of families that is not popular among most working pteridologists. That wouldn’t be such a big deal, but they also lumped together a number of genera in a way that’s similarly unpopular. So our generic boundaries tend to be based on PPG I because, pragmatically, when new literature is published, pteridologists are likely to use those circumscriptions rather than those adopted in POWO. PPG I does not enumerate individual species. However, a popular resource that does so is Hassler & Schmitt’s “World Ferns”: https://worldplants.webarchiv.kit.edu/ferns/ The list there, which usually follows PPG I generic boundaries, feeds the taxonomy at http://pteridoportal.org/portal/ which I am told unofficially reflects “updates” to PPG I.
I have spent a lot of time trying to herd this group of cats! The end goal is to have a taxon framework that completely describes the relationship between our taxa and those in POWO, so that we can see where we match and where we deviate. I don’t know if that’s programmatically accessible, although I would hope it might be. Alas, we are not there yet: it looks like we have about 5,250 taxa in the framework for ferns, and there are about 1,400 (I have it down from 2,500 or so in the past few weeks!) that have not yet been mapped.
My work on this includes compiling errata to be fed to IPNI, POWO and World Ferns, so I can paint a reasonable picture of what changes I expect to happen in the future. Everyone gets (understandably!) tense about molecular phylogeny wreaking havoc on generic arrangements, but happily, I think a lot of that is past. Note that in cases like Asplenium nidus that you mention, there still has to be a morphological diagnosis for valid publication of a new species. e.g., Cystopteris fragilis is a similar complex, but has not been split up because there’s no clear, defining morphological difference to lay hands on. You can’t make new species based solely on DNA sequencing.
There are some distinct areas where we can expect continued instability, which I will lay out below. In general, for big changes in generic circumscription, I try to make a post to the iNat vascular plant curator project and ping top observers/identifiers. I know I will get yelled at if I make big, disruptive changes, so it pays to solicit people’s opinions before putting those things in motion! (It sounds as though changes that affect large numbers of observations need to be coordinated to avoid thrashing the server.) Also, when the sources I’ve mentioned disagree on splitting or lumping a species, I tend to favor splitting (though not inevitably), simply because it’s easier to lump observations in the future then to split them out, especially if ranges overlap.
Lycophytes (clubmosses and allies): PPG I has adopted a fairly “splitty” arrangement here (it definitely leans towards splitting in general). This has been well-accepted in North America, where Flora Novae-Angliae primed us for it, but some of the leading figures in New Zealand feel that it splits too much and plan to adopt a more conservative arrangement. In deference to that, I haven’t attempted to clean up the taxonomy in a number of places. We are agreed to recognize the split of tassel ferns from Huperzia as Phlegmariurus–that may need some ongoing curation.
Marattiales (potato ferns): authorities seem to differ quite a bit in species circumscription in Angiopteris and Archangiopteris–I haven’t sat down to study the problem yet.
Osmundales: PPG I split Osmunda into very small genera and left one hybrid without a valid name. I dislike this particular split and haven’t dealt with it yet. (This involves a change of name to interrupted fern, which will probably raise some eyebrows in North America if carried through.)
Gleicheniales (forked ferns): right now we’re treating Dicranopteris linearis and D. pedata as different species, but I’m not sure there actually is a difference, and a lot of D. linearis IDs may be for the lumped taxon. (This is a weedy and now pan-tropical fern, which is why I mention it.)
Hymenophyllales (filmy ferns): PPG I has taken up Ebihara and Iwatsuki’s division of Trichomanes into several genera, but there has been exception to this from New Zealand, where the characters for splitting have been said not to be useful in the field. Our taxonomy has been left inconsistent here.
Cyatheales (tree ferns and allies): I haven’t really researched this, but there is trouble brewing. Historically (pre-molecular-phylogeny), workers divided between Holttum’s system, which put all of Cyatheaceae in a ~500 spp. Cyathea, and Tryon & Tryon, who split out Alsophila, Sphaeropteris and some other segregate genera. POWO, PPG I, and World Ferns have all recognized the last two genera, and we have some species in them, but many floras adopted Holttum’s system and we have a lot of Cyathea spp. that would need transfer. This needs thoughtful attention before any changes take place.
Lindsaeaceae: New Zealand Flora does not recognize some segregates, considering them weakly differentiated but I think right now that only affects one species of Xyropteris.
Pteridaceae: subfamily Cheilanthoideae is very hard to pick apart and is the subject of ongoing research. It is very hard to impossible to find a set of morphological characters that uniquely circumscribes many genera. POWO tried to deal with this by putting everything in the subfamily in genus Hemionitis but that just led to one big genus that’s not readily circumscribed. Expect more splits here, particularly from Pellaea and Cheilanthes. Some species are being moved to Oeosporangium, which will affect most European “Cheilanthes”, but some combinations are still missing and I think that it’s better to wait a little on that. The vittarioids (shoestring ferns) are also a problem: a proposal is out to conserve Haplopteris over Monogramma but has not yet been voted on, there will be some continued instability there.
Dennstaedtiaceae: I’ve tried proposing a global rearrangement of Pteridium (bracken) but the only feedback I’ve gotten so far is that people dislike their regional taxa being demoted to subspecies. There have also been recent changes in Hypolepis that would affect our New Zealand community; I haven’t checked the state of play there. In general I think Dennstaedtia is subject to further carving up.
Thelypteridaceae: this is a very speciose family. Holttum undertook a massive program of carving it up into genera up to his death in 1990. Some authorities have adopted that. Others have preferred keeping just about everything in Thelypteris (the approach taken by POWO), but this gives a very large and heterogeneous genus; many people taking this approach have recognized subgenera and sections tracking Holttum’s classification. Recent temperate North American floras have recognized seperate genera, but only a few, as it is not very diverse in that region. I expect a major revision to be published with molecular evidence in a year or so dividing it into genera on more rational lines and while I’ve moved a few things around (including the swap that started this), I’ve avoided placing many deviations in the taxon framework in anticipation of future name changes stemming from this.
Blechnaceae: also a point of controversy, with POWO adopting a very large Blechnum, PPG I splitting it a lot, and an intermediate arrangement championed in New Zealand (although some members of the community seem OK with the full splitting). Again, left inconsistent.
Athyriaceae: some workers feel that Anisocampium and Cornopteris should be recognized as segregate genera. Probably a reasonable request. The Athyrium filix-femina complex is a nightmare: we have three taxa in North America which are distinct from the Eurasian type but differences are morphologically subtle and (while combinations exist) would not necessarily be popular to split at species level.
Dryopteridaceae: not sure if Lomagramma should be synonymized with Bolbitis.
Davalliaceae: looks like the current approach is to merge all genera into Davallia. Haven’t really looked into this yet.
Polypodiaceae: it looks as though Phymatopteris (not recognized in PPG I), Paraselliguea and Arthromeris should be merged into Selliguea, following the most recent research. POWO already takes this approach. (Doesn’t affect too many observations.) Subfamily Grammitidoideae was lumped into a very big Grammitis by POWO. Admittedly, this has been a common treatment historically, but there has been a lot of work over the past several decades splitting it up into segregate genera and new combinations will undoubtedly be published there. Fortunately, at least two world experts on grammitids are reasonably active users of iNaturalist. :) Within subfamily Microsoroideae, there is a lot of instability still being resolved. We have some taxa duplicated between Microsorum and Phymatosorus. The latter should probably be merged into Microsorum following PPG , but we can expect continued changes in the future. (A new phylogeny for the subfamily just came out: https://www.sciencedirect.com/science/article/abs/pii/S1055790319303902 and you can see it has a number of holes for genera to be described.)
So thanks to everyone who persevered through this. tl:dr; expect continued instability in Thelypteridaceae, Cheilanthoideae and Microsoroideae as phylogenetic relationships get worked out, tree ferns look like trouble, and we have community disagreements in Lycopodiaceae, Blechnaceae, and Hymenophyllaceae which are not easily resolved.
I hope this provides some reassurance that taxonomic instability is not going to affect every fern you love and hold dear.