In many genuses, species are organized into sections and complexes. When us users add an observation of such a new species, it is not put into the right section and complex. That can easily happen in genuses such as Rubus where the taxonomy is really complicated (subgenus, section, subsection, series). It makes good sense to add the species in bulk in the right categories, because otherwise it leads to disagreeing votes (e.g. with a current section-level ID) when in fact it is just a closer specification.
7 Likes
Personally if I can Iâll add whole genus if genus isnât super big or I gradually finishing those large genera. In Platyhelminhes we have several âsuperuserâ identifiers who can work with those species on very professional level once such species is spawned in their identification queue. It saving their time on flagging and making work more convenient here. Batch adds saving also my time because I know which genera or papers I finished. Also there are no RSS/watch notifications for taxon related flags so I consider batch adds more convenient. In general I recommend to know your community and make own guide based on that. If taxon donât have observers or identifiers itâs probably wasted time but if those positions are occupied it may be worthy in long term basis.
1 Like
It seems like it should be OK to add taxa that are likely to have observations, even if there are none currently.
However, adding many obscure taxa that are not likely to be identified by photo causes errors and additional work for curators. For example, maybe thereâs a genus name change. A user requests the addition of the new genus. Some species are left behind in the old genus, and some new species are added to the new genus, duplicating those in the old genus. Then when the new species are moved, they may need a slight name change to match the gender of the new genus.
This particular problem isnât huge, and it isnât limited to taxa without observations. But there are lots of comparable situations, and it illustrates how additional species require additional labor.
I donât see any major problems adding some species without observations, but there would be a significant amount of manual work (if you can call something as complex as editing a giant database from 10,000 locations around the world while millions of users are accessing it for observations âmanualâ) required to maintain a complete or near complete taxonomy catalog of all life forms. Taxonomy is a moving target, and the more data there is, the more there is to maintain.
From my narrow point of view, there seems to be a pretty good balance of taxa in iNat, along with a good system for adding and maintaining new species. But when adding new species, we shouldnât forget that they will require attention in the future. Itâs exciting to add a newly described species that has or will have observations, or a species with an observation thatâs new to iNaturalist. On the other hand, it may be a lot more efficient to omit a species that has not been seen recently and/or is likely a synonym, at least until there is an observation for it.
10 Likes
Itâs not just additional labor, it also makes the overal taxonomic database larger, meaning search indices need to take all of those nodes into account, etc. So there are technical, database-related downsides to supporting unobserved taxa as well.
8 Likes
This is the main thing I was thinking of when I spoke of âdatabase clutterâ. Thereâs some skepticism above about this, but it makes sense to me.
1 Like
Tony, would âmaximizing the performance of iNaturalistâ or some such be considered as a future subject for a blog post? Thereâs been a distinct shift in guidance over the past several years, much of which (like this) seems to be driven by concerns about site performance and memory usage. Other than a vague notion that places are especially expensive, all of this is very opaque to me, and I would be very interested in knowing what curatorial tasks might help overall site performance.
17 Likes
I wonder if the likelihood of there being observations is influenced by whether the taxon has been added?
5 Likes
As i see it, the addition of âtaxa with no observationsâ can be beneficial in multiple ways. It sounds like there is no question on taxa that have been observed, but iâm absolutely in favour of those somehow deemed likely to be observed. I also see good reasoning for why it can be desirous to add many that supposedly have not been observed yet, but which could be observed.
-
Allows âunsorted observationsâ and those âcurrently at higher ranksâ to have more potential to be refined (e.g. family level observation to species) without any curatorial input (i.e. no need for user to make a flag - wait for response etc, nor import name).
-
Allows users greater awareness of scope of options to make more informed choices about known biodiversity (e.g. if one species is listed, is that genus monotypic or are there potentially far more species?)
-
Allows taxa to be cross-linked with other external databases (e.g. notably wikidata). Such linkage can arguably increase confidence in taxonomic validity (e.g. itâs also listed elsewhere some with details of the relevant literature, history etc), plus facilitates the users directed consultation to that external data (even internally via links e.g. iNat displaying mapped gbif data or wikipedia page)
One of the concerns iâve heard before as above having empty taxa âadds curatorial workâ. Well, what about the workload to answer each user flag for a single taxon addition, then done multiple times? At least in smaller lineages, and especially ones with accessible recent taxonomic studies, then it seems to me logical for a curator to just add the relevant species, form whatever cross-links, then itâs essentially done, right? Of course, above i see it as a valid concern on the need to setup swaps with further revision of those added (the scale of which is surely hard to judge as in the future?), but also need to setup swaps of duplicates (those done by curators often seem to be made by inexperienced, especially not knowing âgoodâ versus poor sources in each taxa). But, that said the system does have some checks in place for some curator actions - at least when trying to add certain names already in the system. That said, the setup could be far easier for curators to see past actions in a taxonomic group though, and what names are already on but inactive, etc!
As for technical âcostsâ of the system indexing and such, iâve inferred that can be a concern (of any database) but never seen any gauge of performance etc to quantify any such vague notions. As for the âcostâ of additional burden of curatorial workload - is anyone else getting paid for curation as iâm certainly not (my point being, what is the âcuratorial costâ that iNaturalist itself is lamenting about - about âhowâ volunteer curators are spending their donated time and effort?). And with my own donated efforts - iâd far rather be making swaps and updates alongside taxonomic revisions than answering a million requests from users to âplease add this one extra speciesâ
7 Likes
my calculation is almost the exact opposite in a lot of these cases!
-
I would much rather be answering âadd taxonâ requests than making swaps. new taxon entries can easily be subsumed; mistakes in committed swaps and taxonomic changes cannot be undone.
-
I agree that it allows higher-level identifications (or disagreements functionalising high-level IDs) to be more refined. however, is this always necessary and informative? just because we can say something is a narrower taxonomic bin does not necessarily add value, especially for those many ranks that have not been verified in the realm of molecular systematics (or other means of support).
-
I am neutral on the issue of âuser awarenessâ. I donât think anyone should be using the availability of names on iNaturalist to assess the validity of names at large. I understand this isnât always how it works, especially for groups that lack good guides or documentation, but thatâs a problem to be addressed via more identification materials, not fixable solely by adding more empty taxon entries and waiting for them to be populated.
-
I am not sanguine enough about external databases to support the idea that a name appearing in both places is appropriate for âtrue confidenceâ, at least specifically in the groups I study (professionally) or know (in an amateur capacity). the first parenthetical category encompasses flowering plants, which I do work on as a molecular systematist; iNaturalistâs main policy (with an exception carved out for ferns) is to try to follow POWO for vascular plant taxonomy, but often as not the direction of information flow goes from iNaturalist flag discussions to Kewâs updates of POWO! on the other hand, for groups like ascomycete fungi without any taxonomic framework, there is no particularly appropriate external database available. for the main group Iâve begun trying to incorporate into my actual work, the powdery mildews, Wikipedia and its ecosystem of wiki resources are abysmally out of date and would be an immense hindrance (or at least a giant sink for curatorial work) if someone attempted to go back and forth between the two. I guess what Iâm saying is that the idea of linking with external resources is nice in theory and kind of works for better-resourced areas of study like birds or certain spots within arthropods, but is questionable to impossible for a large swath of the tree of life.
I completely agree on the first part of the last point though. we probably need a good deal more information about why the âempty branchesâ are costly, and if so how costly exactly.
5 Likes
This whole conversation is an interesting read. Should I just wait for requests to move species to different genera, or add new species, rather than just changing them when they are first published? I had been monitoring taxa on WSC but maybe I shouldnât always do this?
Maybe the balance of pros and cons for adding species works out differently for different types of organisms and different parts of the world.
My experience mostly relates to angiosperms in the Americas, and Iâve seen little reason to limit the addition of new species to those already observed on iNat.
Firstly, I do mostly support the principle of waiting for POWO to add new species. Itâs more efficient to have POWO be our default vascular plant taxonomy source rather than to add species directly to iNat from new publications. To be clear, I donât feel that we need to wait because of the very small chance that a new species may be disputed, just because I donât see a need for us to duplicate the work that POWO is doing.
But plenty of people do want to add newly described species before they show up on POWO and there is rarely any need to obstruct this. In many cases, the authors of the new species are themselves iNat users and have uploaded observations of specimens they used in describing the new taxa.
Given the increasing ubiquity of iNat, itâs fairly likely that observations will be added for many or most newly described species (or discovered hiding with other IDs). I donât see that it helps us to leave these taxa out of iNatâs taxonomy until after the point that we have one or more relevant observations. Adding a taxon now, ideally with an atlas based on known distribution, makes it a lot easier for iNat identifiers to consider it as a candidate ID.
I realize that itâs a challenge to keep iNatâs many functions performing well, and that the number of taxa has some impact on that performance, but I find it hard to believe that adding a species here and a subspecies there based on new publications is going to have a discernable impact. Thatâs entirely different from someone importing a large catalog of unverified names from some dubious (and probably out-of-date) source.
7 Likes
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.