Taxon Concepts as a data model #1852
Comments
|
@dustymc Could you give us your ideas about how this type of model would look in Arctos? |
|
The model I envision is pretty simple - identification_taxonomy (links identifications to names/taxa, which are refined to a source by a collection's preferences) becomes identification_classification - identifications would link to specific classifications/concepts rather than names. We'd have to preserve classification_id and maybe limit how things can change and such, but that's details. Using that with something that can talk to Arctos is pretty simple, if perhaps a lot more labor intensive. (Think #1136, but with maybe-hundreds of options for every name, in addition to the "which name should we use?" thing that Issue's focused on.) Right now, you type "Echidna" into a pick, get one match (and maybe some species and stuff that you don't have enough information to care about, related names, etc.), select it, and the details work themselves out from collection's source preference. In a concept model, you'd type "Echidna" and get (along with all of the species-and-junk, including those in Bitis and Tachyglossus and whatever other synonyms might exists) something that looks very much like http://arctos.database.museum/name/Echidna - I think the full classification in the context of all other classifications is the minimal amount of information you'd need to pick one. Scroll around, find the thing you want, click "use this one," voilà. There are ~14 "concepts" that include Muraenidae on there at the moment, so working out which ones to pick under which situations would be left to ya'll. I'd expect that to grow rapidly (if anyone decides to really embrace this idea, anyway) - any source might include the original publication, then 231 years of publications refining circumscriptions, and publications rejecting publications that tried to refine circumscriptions, and groups of those, and groups excluding certain publications, and field guides, and all the other normal noise. (And for homonyms like Echidna, perhaps the same sort of information for viruses and moths and monotremes and such.) I think that's easily hundreds of "concepts," and some - or most - of them may be different ways of saying the same thing (https://academic.oup.com/sysbio/article/65/4/561/1753624). Accessing that level of complexity with something that cannot talk to Arctos is a great mystery to me. Most specimen data come from spreadsheets and such - things that cannot talk to Arctos. Much/most/all of that is by people who don't even KIND OF have the resources for figure out what definition of eel whoever entered the data a few decades ago might have had in mind. That's "just" a usability problem and there's certainly a way around it. I suspect finding that pathway will involve having the right people in the same room for a few days. Maybe it's as simple as having some sort of default concept (eg, the original description), although I'm not sure how the details of that could work. I think the vast majority of the time we don't think in taxon concepts, so some sort of "just use the name" default might be necessary anyway. ("It's a moose, obviously" - that's all we know or really care about most of the time.) Note that this completely avoids the issue of defining taxon concepts. If you can stuff it into an optionally-ordered key-value array, or stuff some sort of summary into that structure and link to the "real" concept (what we do with WoRMS), you could use it in identifications. |
|
Maybe it's as simple as having some sort of default concept (eg, the
original description), although I'm not sure how the details of that could
work.
Although more complicated I would think it would be better to have such a
default concept be somehow closer in time to the date of the ID. If the ID
was in 2010 and the latest publication to use that name in a taxonomic way
(in a way that qualifies it be considered a taxon concept) is from 2005
then that's the concept that should be default - the most recent.
Original descriptions, esp. old ones, are rarely used or even seen by those
doing the IDs today. If one had to choose a way to be the least wrong most
of the time, it would be to assume the concept used was the most recently
published concept.
…-Derek
On Fri, Jan 11, 2019 at 5:51 PM dustymc ***@***.***> wrote:
The model I envision is pretty simple - identification_taxonomy (links
identifications to names/taxa, which are refined to a source by a
collection's preferences) becomes identification_classification -
identifications would link to specific classifications/concepts rather than
names.
We'd have to preserve classification_id and maybe limit how things can
change and such, but that's details.
Using that with something that can talk to Arctos is pretty simple, if
perhaps a lot more labor intensive. (Think #1136
<#1136>, but with maybe-hundreds
of options for every name, in addition to the "which name should we use?"
thing that Issue's focused on.) Right now, you type "Echidna" into a pick,
get one match (and maybe some species and stuff that you don't have enough
information to care about, related names, etc.), select it, and the details
work themselves out from collection's source preference. In a concept
model, you'd type "Echidna" and get (along with all of the
species-and-junk, including those in Bitis and Tachyglossus and whatever
other synonyms might exists) something that looks very much like
http://arctos.database.museum/name/Echidna - I think the full
classification in the context of all other classifications is the minimal
amount of information you'd need to pick one. Scroll around, find the thing
you want, click "use this one," voilà. There are ~14 "concepts" that
include Muraenidae on there at the moment, so working out which ones to
pick under which situations would be left to ya'll. I'd expect that to grow
rapidly (if anyone decides to really embrace this idea, anyway) - any
source might include the original publication, then 231 years of
publications refining circumscriptions, and publications rejecting
publications that tried to refine circumscriptions, and groups of those,
and groups excluding certain publications, and field guides, and all the
other normal noise. (And for homonyms like Echidna, perhaps the same sort
of information for viruses and moths and monotremes and such.) I think
that's easily hundreds of "concepts," and some - or most - of them may be
different ways of saying the same thing (
https://academic.oup.com/sysbio/article/65/4/561/1753624).
Accessing that level of complexity with something that cannot talk to
Arctos is a great mystery to me. Most specimen data come from spreadsheets
and such - things that cannot talk to Arctos. Much/most/all of that is by
people who don't even KIND OF have the resources for figure out what
definition of eel whoever entered the data a few decades ago might have had
in mind. That's "just" a usability problem and there's certainly a way
around it. I suspect finding that pathway will involve having the right
people in the same room for a few days.
Maybe it's as simple as having some sort of default concept (eg, the
original description), although I'm not sure how the details of that could
work. I think the vast majority of the time we don't think in taxon
concepts, so some sort of "just use the name" default might be necessary
anyway. ("It's a moose, obviously" - that's all we know or really care
about most of the time.)
Note that this completely avoids the issue of defining taxon concepts. If
you can stuff it into an optionally-ordered key-value array, or stuff some
sort of summary into that structure and link to the "real" concept (what we
do with WoRMS), you could use it in identifications.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#1852 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AIraM1_I-lpGuyX95qJqsLQl5a43SXgDks5vCU3CgaJpZM4ZSkes>
.
--
+++++++++++++++++++++++++++++++++++
Derek S. Sikes, Curator of Insects
Professor of Entomology
University of Alaska Museum
1962 Yukon Drive
Fairbanks, AK 99775-6960
dssikes@alaska.edu
phone: 907-474-6278
FAX: 907-474-5469
University of Alaska Museum - search 400,276 digitized arthropod records
http://arctos.database.museum/uam_ento_all
<http://www.uaf.edu/museum/collections/ento/>
+++++++++++++++++++++++++++++++++++
Interested in Alaskan Entomology? Join the Alaska Entomological
Society and / or sign up for the email listserv "Alaska Entomological
Network" at
http://www.akentsoc.org/contact_us <http://www.akentsoc.org/contact.php>
|
|
@DerekSikes I agree, but that involves tracking down ~3 million "most recently published concepts." Maybe "it's just a name, we're not asserting anything" is a better default - although that might require a NULL classification or something equally weird. I think this comes down to what ya'll are willing to do. I'm operating on the assumption that most data entry is going to involve a label/spreadsheet/whatever that just says "Somegenus somespecies" and Curators who are OK with that level of information (eg because they don't have the resources to do anything else). I'd love to require something more specific, I just don't see how we can pull it off. |
|
This is interesting. I like the concept if it could be implemented.
I assume that this way we could have taxon concepts that would be preferred
by different institutions - e.g. Myodes gapperi in a classification that
prefers this version of the genus and the family Cricetidae subfamily
Arvicolinae could be preferred by MSB, and Clethrionomys gapperi in a
classification that prefers the genus Clethrionomys in the family Muridae
would be preferred by MVZ, and Myodes as a beetle could be preferred by an
insect collection If we could display the taxon concept preferences for
each institution when there are more than one possible, then we could
filter these so students doing data entry choose the right one.
And I disagree that collections don't think in taxon concepts already -
they do, just at a broader scale. Perhaps you have "*Somegenus somespecies*" ,
but in the context of a mammal or bird or insect collection. We should at
least know phylum and class.
…On Mon, Jan 14, 2019 at 1:47 PM dustymc ***@***.***> wrote:
@DerekSikes <https://github.com/DerekSikes> I agree, but that involves
tracking down ~3 million "most recently published concepts."
Maybe "it's just a name, we're not asserting anything" is a better default
- although that might require a NULL classification or something equally
weird.
I think this comes down to what ya'll are willing to do. I'm operating on
the assumption that most data entry is going to involve a
label/spreadsheet/whatever that just says "*Somegenus somespecies*" and
Curators who are OK with that level of information (eg because they don't
have the resources to do anything else). I'd love to require something more
specific, I just don't see how we can pull it off.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#1852 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AOH0hDCW5NkmMOULS1AhisA_78ueNw-Pks5vDOzHgaJpZM4ZSkes>
.
|
|
Implemented is easy. Used - maybe not so much... There is no "preferred" in the model. Maybe we can figure out some default or something as above, but the "core" would be explicitly picking a concept (unless someone comes up with something clever...). I think it would be generally finer-grained than you've described. You'd have M. gapperi (limited to WHATEVER because SOME PUBLICATION or something - maybe something about range or morphology or DNA or karyotype or song or parasites or ...), and M. gapperi (potentially with exactly the same hierarchy) limited to SOMETHING ELSE because SOME OTHER PUBLICATION (and maybe hundreds of other concepts, potentially all with exactly the same hierarchy). Knowing phylum or class won't help very much - that'll get rid of Myodes-the-bug, but still (probably) won't get you to the single classification/concept you need to create an ID. You're probably right that we do think in "fuzzy taxon concepts," and Derek is probably right in that those aren't really THAT fuzzy, it's just that we don't record anything useful so someone in the future has to guess who's idea of M. rutilus was used, and how close that concept was adhered to, when the ID was applied. |
Wouldn't each concept have a publication date associated with it? Even just a year? Then you just pick the date closest to today's date? |
|
It sounds like we agree that this is the way to go, but we need to get together in a room and work out the details. I would suggest we do this at SPNHC, but I think I would rather make this a meeting about just one thing without the distractions of of other presentations and ideas. Thoughts? Who really wants to be included in the in-person meeting? |
No, that would take us into defining taxon concepts. And the "closest to today" 'concept' might be a list of publications all refuting your publications or something.... |
Jegelewicz
assigned campmlc, DerekSikes, acdoll, sharpphyl, anna-chinn, mbprondzinski, Jegelewicz and dperriguey
|
To make this work we would need a massive publications database that holds
all the concepts we'd want to use. If we can't import that we'd have to
have users create these on the fly as they need them.
I could see the following:
For the majority of name uses in Arctos it will not be clear what the
concept was. I could see 2 options for these cases:
1) taxonname with concept = 'unknown'
2) taxonname with concept = 'most recent publication that defines that
name' with a qualifier field stating the confidence of the association, in
this case = 'likely'
For those who want to use taxon concepts and know the concept they're
using, would opt for
3) taxonname with concept ='whateverpublication the concept came from' with
a qualifier field stating the confidence of the association, in this case =
'certain'
I suppose for the #1 option above the qualifier could be set to 'certain'
since this would be true - the user is certain they don't know what concept
was applied.
thoughts?
What outstanding questions remain on this?
…-Derek
On Wed, Jan 16, 2019 at 1:41 PM Teresa Mayfield-Meyer < ***@***.***> wrote:
Assigned #1852 <#1852> to
@DerekSikes <https://github.com/DerekSikes>.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
<#1852 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AIraMwNX5CymTRmkRO7nrA7hNEozk985ks5vD6p9gaJpZM4ZSkes>
.
--
+++++++++++++++++++++++++++++++++++
Derek S. Sikes, Curator of Insects
Professor of Entomology
University of Alaska Museum
1962 Yukon Drive
Fairbanks, AK 99775-6960
dssikes@alaska.edu
phone: 907-474-6278
FAX: 907-474-5469
University of Alaska Museum - search 400,276 digitized arthropod records
http://arctos.database.museum/uam_ento_all
<http://www.uaf.edu/museum/collections/ento/>
+++++++++++++++++++++++++++++++++++
Interested in Alaskan Entomology? Join the Alaska Entomological
Society and / or sign up for the email listserv "Alaska Entomological
Network" at
http://www.akentsoc.org/contact_us <http://www.akentsoc.org/contact.php>
|
Only if you want to base your concepts off of publications. "Whatever WoRMS does with aphiaid=12345" is a perfectly valid (if mostly useless...) "concept" in the model I'm proposing, for example.
It wouldn't necessarily be completely unknown - the default would (hopefully) be something like we do now, eg http://arctos.database.museum/name/Sorex%20vagrans#Arctos. We're definitely not talking about Sorex vagrans (the jellyfish) because we have class and such in there, but there's not much detail either.
Or group of publications, or most recent not including THAT publication, or ...
That's identification - you're not qualifying the taxon in any way, just how the specimen fits in to it.
I don't think that's quite accurate. They're certain (or not) that it's a moose, they're just applying some waffly definition/circumscription/whatever - concept - of "moose."
I think just how we use it, if nobody can find holes in the idea of replacing... taxon_name_id (foreign key-->taxon_name.taxon_name_id) in that with a foreign key to taxon_term.classification_id (and we can figure out how to preserve that identifier and still use the hierarchical editor and have enough processors to do stuff with this and all that jazz). |
|
https://www.loc.gov/standards/sourcelist/index.html |
|
HMMMMMMM...this is interesting. Perhaps we should have "arctos" in https://www.loc.gov/standards/sourcelist/taxonomic.html How many of these could provide WoRMS-like data to our taxonomy? |
One purpose of GlobalNames is to make the answer, "who cares?" With GN, we write to one API and get everything they have. Without that abstraction, we'd need to write to 182ish (https://resolver.globalnames.org/data_sources) APIs. There's no obvious standardization among those sources - I doubt much code could be reused. Re-creating what GN provides would require a tremendous amount of resources. For whatever reason GN doesn't contain all of the information from WoRMS and isn't updated very often, we couldn't get them (or WoRMS, or something) to fix that, there's some additional complexity (we need specific classifications in some cases), so we wrote code to WoRMS. We can do that for other sources too if there's some compelling reason (eg, someone's going to use it to catalog), but for most things I think the best path is to encourage the source and GN to deal with it. |
I have always seen Arctos more as a consumer than an authority, even though that's not entirely where we've found ourselves. I would like to see more WoRMS-like connections, and less local editing/"authority building." Ideally Curators (and/or their representatives) who want to could put their taxonomist hat on, log in to something like WoRMS, make changes there, and see them magically appear in Arctos. Curators who don't want to wear that hat could just passively use data from some source, or if we go to some more concept-like model from any source by picking individual "classifications" from the local cache. Even more ideally, GN would become more active in keeping things complete and current and we'd just maintain one API to do that. I don't know how realistic any of that is, but I do think it's the best model for everyone involved. |
|
for review and discussion: http://ubio.org/ @dustymc promising? |
|
uBio initially intended to implement the Ballew thesaurus, and we spent a lot of time talking to them before they started writing code. If they'd have done what they set out to do, we'd likely just use them. (Or they'd have killed us all by sucking up every electron on the planet trying to build a transitive closure table....) What they actually implemented is a "curated view" much like ITIS and everyone else, which is not very useful to Arctos as a whole. It could be useful to individual collections. I don't think there's anything that could remotely be described as taxon concepts in the data they have, but I haven't looked closely in a long time and I could be missing something. |
|
Check out the search for Diplura in their Nomenclator Zoologicus: http://ubio.org/NZ/search.php?search=Diplura&authority=&category=&publication=&year=&comments=&exact=1&advanced=1&vol=&page= |
|
thoughts on usability
Default action from the specimen bulkloader (and other string-based tools) would be to use the classification/concept within the collection's preferred source which does not have a taxon_concept_id. (And nothing does now, so this would essentially be no changes that a user would notice.) The specimen bulkloader could be modified to somehow accept taxon_concept_id - eg, Some species (someID) would use the concept under Some species with an ID of someID, and error if that doesn't exist or if there are multiple matches. taxon_concept_id would be a string so "Conus Linnaeus, 1758" or "http://www.marinespecies.org/aphia.php?p=taxdetails&id=137813" would be acceptable values. There is not a 1:1 name-concept relationship, so a unique key would be difficult or impossible - it would likely be possible to create many concepts "named" Conus Linnaeus, 1758 under one taxon, in which case references using the name+ID would be ambiguous so would error. I think this requires careful users and good documentation. For internal forms, we can pass around data objects and none of the above is a concern. Would anyone want to load specimens with concepts? (Seems inevitable...) It would be possible to link identifications to concepts outside the collection's preferred source - it would be possible to eg, use a concept under WoRMS (via Arctos) (or anything else) for a collection which generally prefers the Arctos Plants (or whatever) source. Perhaps concepts should even be managed in their own source(s), which would keep "normal" sources cleaner/easier to manage. E.g., that could be exploited to disallow someone adding taxon_concept_id to a "default" classification. Used concepts cannot change; "changes" create new concepts. It would be exceptionally useful to have a consistent "backbone" in support of search. This doesn't have to be a Source (or sources - perhaps it's easier to manage at phylum/kingdom/etc.) that anyone uses, it would just facilitate search to compensate for (perhaps purposefully) inconsistent data with "concepts." This could be managed as a hierarchy and excluded from the single-record editor. Concepts may be purposefully inconsistent and will be managed singly; it's likely safe to proceed with change requests to the single-record editor. Example: http://dx.doi.org/10.1093/zoolinnean/zlx040 split a taxon/created a new concept and could serve as a most-basic test case. |
This and other comments keep bringing me back to the "preferred by" solution. Perhaps the easiest thing would be that any time there is more than one classification in a source, collections using that source are notified and can select the classification they wish to use for all identifications in their collection. We could report this in the Low Quality Data section where collections could find a list of all the taxa for which there are more than one classification but they have not selected a preference. This would take the decision about taxonomy out of the hands of students entering data. I think we would want to start off making whatever classification a collection is using right now their preference, so that when Derek comes along and starts adding classifications no one will suddenly need to choose preferred classifications for 100's of names. Otherwise, collections who don't choose a preference will end up with what we have now - a mash-up of all classifications associated with the name. This means their stuff will be found (although sometimes in error). The main challenge I see to this is how do we record preference and ensure that someone can't go around changing the preferences for someone else's collection? I don't think throwing it into the Classification Metadata would work, but maybe Dusty could fix it so that I can only add/delete preference for collections to which I have access. Also, there will be the need to track changes in the preferred classification. Somehow, when I decide to change from one preference to another, I would like to create a notation in the identification section of the specimen record. |
If that's the only objective, the model we are currently in accomplishes it much better than taxon concepts could. If you're trying to be more precise than taxonomy allows in identifications, you might NEED taxon concepts. If you're trying to sort beetles from mice, you're almost certainly going to absolutely hate the extra workload and I never see the benefits.
This also seems to suggest that we don't need a taxon concept model. What's the point if the people who do most of the work can't access the complexity??
I don't think taxon concepts can change and remain anything recognizable as a taxon concept. I think we have to find some sort of "default" or "preferred" to be able to use this, but that's just help in selecting a concept when you have no preference - it's procedure, not data. Changing the preference could not change existing data. |
|
There's an attempt at a summary here: https://docs.google.com/document/d/1tR7FOYd_XCLl4o_YrKdUdkwmZu3Fun3dv0xlawyW1zA/edit?usp=sharing |
Exactly. What I am suggesting is that we keep the model we have. This way, data entry is easy - students (or anyone entering data) just have to pick the name. The change I suggest is that as long as only one classification exists in any taxonomy "source", everyone uses it. Let's take Diplura, which in GBIF includes:
If all of these classifications were included in source Arctos, then anything identified as Diplura would get a crazy mash-up of higher taxa. I propose that COLLECTIONS should be able to tag one of these as preferred so that only that classification will be applied to Diplura in the associated collection. Again, this will not solve Derek's issue and isn't really using "taxon concepts" BUT it does allow us to maintain homonyms in a single taxonomy source. I see it as a baby step. |
|
If you're suggesting that CollectionA can pick Diplura-the-spider for one specimen and Diplura-the-butterfly for another, then the only way I see to do that is #1852 (comment) - move the pointer from taxonomy to classifications. It's a taxon concept model, even if the concepts are flaky. (And no model excludes the possibility of flaky data.) If you're suggesting CollectionA has to preemptively go say "All our Diplura are spiders" then this just looks like a really complicated way to split classification sources. It also obfuscates the path between names and "my" classification, unless the "this is mine" bit lives in the classification itself or something. I don't think I'm quite understanding something. |
|
move the pointer from taxonomy to classifications.
yes, but of course the lowest name in the classification is the name in
taxonomy (ie merge the two).
…-Derek
On Tue, Apr 16, 2019 at 1:59 PM dustymc ***@***.***> wrote:
If you're suggesting that CollectionA can pick Diplura-the-spider for one
specimen and Diplura-the-butterfly for another, then the only way I see to
do that is #1852 (comment)
<#1852 (comment)> -
move the pointer from taxonomy to classifications. It's a taxon concept
model, even if the concepts are flaky. (And no model excludes the
possibility of flaky data.)
If you're suggesting CollectionA has to preemptively go say "All our
Diplura are spiders" then this just looks like a really complicated way to
split classification sources. It also obfuscates the path between names and
"my" classification, unless the "this is mine" bit lives in the
classification itself or something.
I don't think I'm quite understanding something.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#1852 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AIraM97tKaA3vAYxW5j8M3GGNb4vE_S3ks5vhkexgaJpZM4ZSkes>
.
--
+++++++++++++++++++++++++++++++++++
Derek S. Sikes, Curator of Insects
Professor of Entomology
University of Alaska Museum
1962 Yukon Drive
Fairbanks, AK 99775-6960
dssikes@alaska.edu
phone: 907-474-6278
FAX: 907-474-5469
University of Alaska Museum - search 400,276 digitized arthropod records
http://arctos.database.museum/uam_ento_all
<http://www.uaf.edu/museum/collections/ento/>
+++++++++++++++++++++++++++++++++++
Interested in Alaskan Entomology? Join the Alaska Entomological
Society and / or sign up for the email listserv "Alaska Entomological
Network" at
http://www.akentsoc.org/contact_us <http://www.akentsoc.org/contact.php>
|
That's not a requirement, although I can't think of a reason it shouldn't be true.
Explain please. |
|
merge as in combine rather than replace. Eliminate redundancy, if it
exists, combine all elements unique to either into one table classification
table.
…-Derek
On Tue, Apr 16, 2019 at 2:53 PM dustymc ***@***.***> wrote:
the lowest name in the classification is the name in taxonomy
That's not a requirement, although I can't think of a reason it shouldn't
be true.
merge the two
Explain please.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#1852 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AIraM8LJpCG-J8ZUxZ8dLiN_Epq-Z1Nmks5vhlSCgaJpZM4ZSkes>
.
--
+++++++++++++++++++++++++++++++++++
Derek S. Sikes, Curator of Insects
Professor of Entomology
University of Alaska Museum
1962 Yukon Drive
Fairbanks, AK 99775-6960
dssikes@alaska.edu
phone: 907-474-6278
FAX: 907-474-5469
University of Alaska Museum - search 400,276 digitized arthropod records
http://arctos.database.museum/uam_ento_all
<http://www.uaf.edu/museum/collections/ento/>
+++++++++++++++++++++++++++++++++++
Interested in Alaskan Entomology? Join the Alaska Entomological
Society and / or sign up for the email listserv "Alaska Entomological
Network" at
http://www.akentsoc.org/contact_us <http://www.akentsoc.org/contact.php>
|
|
So basically taxon concepts (identifications<-->classifications), but without the authoritative "anchor" tieing related stuff together? And I found a classification (1067358 of them, actually...) that doesn't end with the name: http://arctos.database.museum/name/Poecilophis#ArctosRelationships |
|
Could we have the option of searching on (or entering) the display name or the name string which include the author instead of just the taxon name? display_name: Poecilophis Kaup, 1856 Or add to the taxon status "hemihomonym" and "homonym" and for those allow the addition of author to differentiate? Wouldn't deal with everything but would take care of most of what I see. |
Yea, if it's in a classification somewhere I can search it.
You're going to have to be a LOT more specific before I can answer that.
I don't have any objections, but that sounds like a lot of work to make redundant data.
If you mean to the namestring, #1803 with a built-in extra-randomizer does not sounds like fun to me.
PLEASE, elaborate. What is it that we're trying to take care of? |
Nope - suggesting collection A picks one or the other and sticks with it.
I am trying to avoid very single collection having it's own source, which defeats the purpose of a collaborative system. And yes, I assumed the "this is mine" thing would live in the classification metadata. I thought I knew what we were doing with "taxon concepts" but every time I think I know, someone says something that makes me think it will not work. I have been looking for something to help us deal with multiple classifications related to a single name when they occur. We have talked in circles about this for over a year and are getting nowhere - perhaps we need some fresh viewpoints. |
|
@Jegelewicz @campmlc to pick a collection and work with Dusty to create a new taxonomy source that pulls from Arctos. |
Avoiding Taxon ConceptsThere are significant usability issues surrounding taxon concepts - it's just more complex data, so it's more difficult to use in most every way. The model generally seems like overkill for the kinds of problems we're trying to solve. Most of those problems involve homonyms, and there is a reluctance to split classifications in order to share data/updates. Potential not-concepts solution: create "dynamic" sources which are based on collection-defined criteria and auto-refresh themselves periodically. Selection could cross sources, include things like taxon_status or various ranks, etc. Data would be managed in the shared (eg, "Arctos") Source(s) and the dynamic source would be refreshed from updates. Outstanding questions and concerns:I still don't have an example of an actual problem. I can think of two potentials:
https://arctos.database.museum/name/Diplura comes up from time to time, but at least the "Arctos" data are likely just wrong - surely the term isn't actually both a class and order for the same individuals? #1936 (and similar) - we are aggressively pushing things that are very likely to cause problems into shared classifications. I don't think there's anything to share between eg, taxa used by a bird collection and taxa used by a nautiloid collection; those taxa are created by very different user groups, and are probably best managed by different user groups. There is much more to share between eg a modern mammal collection and a paleo collection cataloging lots of Pleistocene material. I'm not sure where to draw any lines, but I suspect there are some in there and we're pushing them in directions that create unnecessary work. Would anyone use taxon concepts as a way to disambiguate taxa (not names) in a way that can't be accomplished with "ID sensu"? If so, perhaps we should figure out how to mitigate the usability issues. If not, perhaps an alternative approach makes sense until we're forced into the more complex/precise model. |
Cam Webb |
|
Notes from meeting: https://docs.google.com/document/d/19cbpGwfQJ52mt89fCag5VU-kh2Q6wuEddzpD_zw1BvE/edit# Cam's plan is to have an additional identification field linked to taxon concepts. He will send Dusty some data and Dusty will evaluate for implementation. Adding to Taxonomy Committee Agenda. |
|
Cam Let's do this! Committee says let's set it up. |
|
Did I miss the meeting?! Crap!
|
|
Closing this as dupe of #2267 |
A thread for exploring the idea.
Related issues/comments include
#1136
#1809
#1817
#735 (comment)
#912 (comment)
#1803 (comment)
#1805 (comment)
#983 (comment)
#1609 (comment)
The text was updated successfully, but these errors were encountered: