Arctos: New Collecting Source

[Jegelewicz]

Issue Documentation is http://handbook.arctosdb.org/how_to/How-to-Use-Issues-in-Arctos.html

Table
http://arctos.database.museum/info/ctDocumentation.cfm?table=CTCOLLECTING_SOURCE

Value
exhumed

Definition
Dug out from the ground.

Collection type
N/A

Attribute data type
N/A

Attribute value

N/A

Attribute units
N/A

Part tissue flag
N/A

Other ID BaseURL
N/A

Context
"wild caught" not appropriate for paleo collections

Priority
I would like to have this resolved by date: 2019-02-28

[mbprondzinski]

This works for me.

[dustymc]

What are we trying to do here?

Unless there's some reason not to or I'm missing something significant, I think I'd prefer one term that means something like "got there by natural processes; useful for range maps" rather than one for plants and one for fossils and one for ....

[Jegelewicz]

@dustymc I thought this is what you were suggesting....

https://github.com/ArctosDB/arctos/issues/1912#issuecomment-466066408

[dustymc]

@Jegelewicz it is, I think, I'd just prefer one term to cover all instances of "got there by natural processes; useful for range maps." "Exhumed" doesn't work for wild-caught animals (except maybe gophers?!?), wild-caught is weird for plants, fossils, cultural objects, etc. I've seen "found in nature" thrown around - same idea as "wild caught," but still awkward for a lot of Arctos.

[Jegelewicz]

OH, got it, so we would replace "wild caught" with something else? How about

removed from nature

Found or taken from a natural habitat. Examples include animals caught in the wild, plants or plant parts not cultivated by people, and fossil or mineral material exhumed from substrate where they were deposited.

[dustymc]

I think we're all on the same page, but that seems to conflict with https://arctos.database.museum/info/ctDocumentation.cfm?table=CTSPECIMEN_EVENT_TYPE (eg, "not killed or removed").

Term: ?????

Definition: Found or taken from a natural habitat or expected environment [ @AJLinn ??]. Examples include animals caught in the wild, plants or plant parts not cultivated by people, and fossil or mineral material exhumed from substrate where they were deposited. Likely considered a valid range or distribution point.

[campmlc]

This has always been an awkward code table because it doesn't work for
cultural collections either. I think maybe we need to back up and consider
allowing different vocabulary for different collections, but have the
vocabulary itself have a code table that indicates whether it should map or
not. Eg, keep wild caught and captive, add in terms appropriate for Paleo
and cultural collections.

On Thu, Feb 21, 2019, 2:35 PM dustymc <[email protected] wrote:

I think we're all on the same page, but that seems to conflict with
https://arctos.database.museum/info/ctDocumentation.cfm?table=CTSPECIMEN_EVENT_TYPE
(eg, "not killed or removed").

Term: ?????

Definition: Found or taken from a natural habitat or expected
environment [ @AJLinn https://github.com/AJLinn ??]. Examples include
animals caught in the wild, plants or plant parts not cultivated by people,
and fossil or mineral material exhumed from substrate where they were
deposited. Likely considered a valid range or distribution point.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-466176195,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AOH0hOEdJh3wBhOuUx7uJ5ZjebQuOcfmks5vPxEdgaJpZM4bH33n
.

[dustymc]

different vocabulary for different collections

Collections are arbitrary. http://arctos.database.museum/guid/UAM:Mamm:63998 could have been cataloged by archeologists, and should filter out with all the other sea cow bits.

Or maybe not - maybe it should filter out with all the other worked sea cow bits and we need a "probably died nearby, but possibly transported by humans" value in addition to a way of saying "wild-caught" that works for plants and dinosaurs??

[campmlc]

Yes, but it doesn't work the other way.
A basket: https://arctos.database.museum/guid/UAM:Arc:0204-2076 is not
"wild-caught" or "captive".
Is it "collected from nature"? Maybe not - does an archaeological site
count?

Maybe we just need to move all this terminology to specimen events and
attributes, and instead have Collecting Source be "mappable" vs "not
mappable"?

On Thu, Feb 21, 2019 at 3:06 PM dustymc [e-mail] wrote:

different vocabulary for different collections

Collections are arbitrary.
http://arctos.database.museum/guid/UAM:Mamm:63998 could have been
cataloged by archeologists, and should filter out with all the other sea
cow bits.

Or maybe not - maybe it should filter out with all the other worked sea
cow bits and we need a "probably died nearby, but possibly transported by
humans" value in addition to a way of saying "wild-caught" that works for
plants and dinosaurs??

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-466186356,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AOH0hL5k-dQwm2L121Dc00CecM9n4aXUks5vPxh0gaJpZM4bH33n
.

[KatherineLAnderson]

For fossil collections, "in situ" (it's found where it was deposited) versus "ex situ" (it's found in this place but it was moved there _recently_ by gravity, water, other act of nature) might work. I don't know how it would apply to other collections... probably poorly. I personally like wild caught or a broad term that can apply to everything.

[dustymc]

doesn't work the other way.

Sure it does - cultural collections are tremendous and underutilized resources of biological material, and they're a lot more valuable for that if we can say something about how that material came to be at various events. If that basket is made out of skin, was it caught locally or bought on ebay?

need to move all this terminology to specimen events

That's where it is???

attributes

In what sense?

"mappable" vs "not mappable"

I think that's the core of the Issue, but it would be useful (especially for cultural stuff) to be more specific. We should have the potential to say both "seals naturally occur here" and "people who live(d) here use(d) seal-bits."

[dustymc]

"in situ" versus "ex situ"

Maybe?!?

[Jegelewicz]

"in situ" versus "ex situ"

I really like those terms - the only problem is they aren't "expected" by everyone and both of those are essentially "wild caught" from the "mapping" standpoint.

[Jegelewicz]

Term: wild caught

Definition: Found or taken from a natural habitat or expected environment. Examples include animals caught in the wild, plants or plant parts not cultivated by people, and fossil or mineral material exhumed from substrate where they were deposited. Likely considered a valid range or distribution point.

Maybe it is just the definition that needs work here....or maybe I need to sleep on it.

[dustymc]

aren't "expected"

Neither are "wild caught" dinosaurs!

I think https://terms.tdwg.org/wiki/dwc:establishmentMeans is the closest DWC (=expected?) term. I dislike the waffly/nonexistent lines between introduced/naturalised/invasive/managed.

in situ==>where it's supposed to be; useful for things like range maps
ex situ==>out of position; not useful for things like range maps

IDK, my Latin is even worse than my English, but these look like a possibility to me.

[AJLinn]

If that basket is made out of skin, was it caught locally or bought on ebay?

We simply don't use this field, nor habitat. Our objects rarely are made from a single source of material - they are what are referred to as "composite objects" - things made from 3 or more materials. The way the maker acquired the materials are going to be just as complex... sea otter could have been purchased from a furrier, beaver was trapped locally, reindeer skin was traded from Siberia, sinew was collected from a reindeer they hunted from a herd that comes thru their area once each year, textile lining was salvaged from a sugar sack they got from a village store... and so on.

I think that you should not worry about cultural collections when you consider this field. Archaeology might have a different position on this though... @sjshirar ?

[Jegelewicz]

in situ==>where it's supposed to be; useful for things like range maps
ex situ==>out of position; not useful for things like range maps

If we use the terms that way - it's perfect. But seems like it might confuse paleo people who expect:

"in situ" (it's found where it was deposited)
"ex situ" (it's found in this place but it was moved there _recently_ by gravity, water, other act of nature)

Which are both "in situ" per the definitions you propose.

[KatherineLAnderson]

aren't "expected"

Neither are "wild caught" dinosaurs!

I think https://terms.tdwg.org/wiki/dwc:establishmentMeans is the closest DWC (=expected?) term. I dislike the waffly/nonexistent lines between introduced/naturalised/invasive/managed.

in situ==>where it's supposed to be; useful for things like range maps
ex situ==>out of position; not useful for things like range maps

IDK, my Latin is even worse than my English, but these look like a possibility to me.

ex-situ can be it fell a couple feet out of the outcrop via gravity, and it can still be used for range maps. Maybe this is more detailed than we need to be for this field?

[dustymc]

Which are both "in situ" per the definitions you propose.

I was going to say something about the intent of my definition, but...

couple feet

That's why I don't like outright saying "not suitable for maps" - the suitability for purpose depends on things like the resolution of the map. Maybe we need some sort of finer-scaled flag/more terms/???????? for float and salamanders in the parking lot and such?? We should somehow be able to get at both "mammoths from {county}" and "we're not sure what formation this diatom fell out of."

[campmlc]

I like
in situ==>where it's supposed to be; useful for things like range maps
ex situ==>out of position; not useful for things like range maps
and if not that, then
"mappable" vs 'not-mappable" and leave it at that. Let the decision be made
at the individual specimen level.

On Thu, Feb 21, 2019 at 4:41 PM dustymc [e-mail] wrote:

Which are both "in situ" per the definitions you propose.

I was going to say something about the intent of my definition, but...

couple feet

That's why I don't like outright saying "not suitable for maps" - the
suitability for purpose depends on things like the resolution of the map.
Maybe we need some sort of finer-scaled flag/more terms/???????? for float
and salamanders in the parking lot and such?? We should somehow be able to
get at both "mammoths from {county}" and "we're not sure what formation
this diatom fell out of."

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-466214425,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AOH0hG5NWTM-sBjQ_Z7I_QswjSBAQOJbks5vPy6mgaJpZM4bH33n
.

[Jegelewicz]

"mappable" vs 'not-mappable" and leave it at that. Let the decision be made at the individual specimen level.

Wait, are you suggesting we make it simple? :-)

Do we believe that everyone understands that "wild caught" and "captive" DO anything (and how would anyone outside of Arctos know what "mappable" and "un-mappable" mean, for that matter, how does anyone inside of Arctos know if what we think is mappable will be mappable to anyone else...)? Should they DO anything or should they just be information for users of the data to sort out. As it is, we are offering multiple accepted "collection" events for a whole bunch of stuff. Anyone who wants to use the information will need to sort through that to decide which event best represents what actually occurred. Maybe we are just trying too hard to DO stuff.

[dustymc]

https://github.com/ArctosDB/arctos/issues/1942#issuecomment-466214425

I think much of the time we're just not going to have this data, but when we do we should have a place for it. We can't force external users to consider all aspects of our data, but we can and should make it possible for them to exclude San Diego Zoo ringed seals from large-scale range maps or (should we ever get sufficient resolution in the data) parking-lot salamanders from fine-scale habitat maps.

[campmlc]

I agree.
As an aside, if we decide to go with in-situ /ex-situ (any objections?), then I'd like to move "captive" to an attribute. We already have "experimental" as an attribute in MSB:Host to deal with mice that were infected with tapeworms etc. Noting that an animal was a captive affects its other attributes such as weight, reproductive status, and ultimately will affect isotopic signatures etc. so this should be recorded.

[Jegelewicz]

Let's bring the change from wild caught to in situ and captive to ex situ to the AWG on the 7th

as part of the transition, anything "captive" will be assigned an attribute:

captive: organism was either raised in captivity or held captive for a period of time between capture in the wild and preservation

with

attribute date = collection event date (began or ended?)
attribute assigned by agent = collector agent or specimen event assigning agent?
attribute remark = auto assigned due to change in collecting source terminology

[dustymc]

I think I like, but with some adjustment.

Let's change "captive" to something more generic ("establishment means"?) and throw up another code table to control the values. The event just says "probably didn't get here without help" ("ex situ") and we can say nothing else (no attribute), or "captive," "float," "cultivated" (not sure how that's not 'captive'), or whatever, or any combination of those, all with a dedicated place (attribute remarks/method) to say WHY and WHO thinks that. "Experimental" should fit in there as well (alongside "captive" if necessary). I think that starts to get at coarse-grained vs. fine-grained suitability for use.

That's not quite modeled right and maybe we'll need to revisit if we get a lot more of these kinds of data, but it looks like it's reasonably capable of holding most of the stuff we're immediately likely to know.

[dustymc]

I think this is where we are:

Data

http://arctos.database.museum/info/ctDocumentation.cfm?table=CTCOLLECTING_SOURCE

gets two new terms:

• in situ==>where it's supposed to be; useful for things like range maps
• ex situ==>out of position; not useful for things like range maps

better definitions appreciated.

New attribute "establishment means" definition "Refinement of collecting_source." linked to new code table ctestablishment_means with values

• captive --->Specimen was taken from captivity.
• experimental-->organism and its host do not represent a natural occurrence

We can add values as needed, to refine especially ex situ to "4 inches below the formation from which it probably came" or "purchased on ebay, probably from Europe" or whatever.

Migration Path

• COLLECTING_SOURCE=field photo-->COLLECTING_SOURCE in situ
• COLLECTING_SOURCE=museum photo--> @DerekSikes what should I do with http://arctos.database.museum/SpecimenResults.cfm?collecting_source=museum%20photo
• COLLECTING_SOURCE=unknown-->COLLECTING_SOURCE ????????????????? Maybe we need "unknown" to survive this migration; it's a required field in specimen_event.
• COLLECTING_SOURCE=wild caught-->COLLECTING_SOURCE in situ
• COLLECTING_SOURCE=captive-->COLLECTING_SOURCE ex situ + add attribute establishment means=captive
• attribute experimental value='yes'--->attribute establishment means value='experimental'
• attribute experimental value='no'--->??? @campmlc what is the intention of these data? Isn't everything not marked as such non-experimental==this can just go away???

Cleanup

• delete attribute 'experimental' (pending a resolution of the 'no' values)
• delete COLLECTING_SOURCE=field photo
• delete COLLECTING_SOURCE=museum photo
• delete COLLECTING_SOURCE=unknown (or not, see above)
• delete COLLECTING_SOURCE=wild caught

check code for places collecting_source is used, fix as necessary

[Jegelewicz]

in situ==>where it's supposed to be; useful for things like range maps
ex situ==>out of position; not useful for things like range maps

better definitions appreciated.

in situ==>inside, on site, or in the natural position in time and space
ex situ==>outside, off site, or away from the natural position in time and space

borrowed from: http://www.oxfordreference.com/view/10.1093/oi/authority.20110810104854504

New attribute "establishment means" definition "Refinement of collecting_source." linked to new code table ctestablishment_means with values

captive --->taken from captivity
experimental-->organism and its host do not represent a natural occurrence

I would add

purchased --->acquired through the payment of money or its equivalent
cultivated --->raised or grown with human skill *maybe the same as captive? Plant people would love this though....*
introduced --->not native to a specific location
drift --->deposited by natural agents such as wind, water, or geologic shifting

COLLECTING_SOURCE=unknown-->COLLECTING_SOURCE ????????????????? Maybe we need "unknown" to survive this migration; it's a required field in specimen_event.

I think we should almost always have an option for "I don't know" so we should leave "unknown"

delete attribute 'experimental' (pending a resolution of the 'no' values)
delete COLLECTING_SOURCE=field photo
delete COLLECTING_SOURCE=museum photo
delete COLLECTING_SOURCE=unknown (or not, see above)
delete COLLECTING_SOURCE=wild caught

Also delete COLLECTING_SOURCE=captive?

[dustymc]

Thanks! I'll go with your definitions if nobody has anything better.

I would add...

It's just a code table, so not much problem.

I don't really want to bog this down on cultivated/introduced/etc., but as "secondary" data this may be a good place to get at those fine-scale distinctions without mucking up my auto-generated "map of Whatever County [non]natives" (for which I can just use collecting_source).

Yes I think we may be stuck with "unknown" in source. For the attribute we just don't include it when we have nothing to say.

Yes the plan is to delete everything except in|ex situ.

[Nicole-Ridgwell-NMMNHS]

I really like this solution.

[campmlc]

Can we autocreate an attribute of captive for all captive sources that are
going away, please?

On Wed, Mar 20, 2019, 12:45 PM Nicole-Ridgwell-NMMNHS <
[email protected]> wrote:

I really like this solution.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-474924042,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AOH0hMO2zGivZxZR5JeM_Sdeo9V-Q9dNks5vYmWtgaJpZM4bH33n
.

[dustymc]

@campmlc the plan is in https://github.com/ArctosDB/arctos/issues/1942#issuecomment-473356057 (so yes) - and can you comment on 'not experimental' from the same place?

[campmlc]

attribute experimental value='no'--->??? @campmlc what is the intention of these data? Isn't everything not marked as such non-experimental==this can just go away???
Exactly. There is no need to autocreate a NO value, only a YES value.

[dustymc]

I think this is ready to go, except museum photo for which I'll do something random and throw that in some remarks field somewhere if it's still hanging around. Here's the revised plan:

Data

http://arctos.database.museum/info/ctDocumentation.cfm?table=CTCOLLECTING_SOURCE

gets two new terms:

• in situ==>inside, on site, or in the natural position in time and space
• ex situ==>outside, off site, or away from the natural position in time and space

New attribute "establishment means" definition "Refinement of collecting_source." linked to new code table ctestablishment_means with values

• captive --->Specimen was taken from captivity.
• experimental-->organism and its host do not represent a natural occurrence

Migration Path

• COLLECTING_SOURCE=field photo-->COLLECTING_SOURCE in situ
• COLLECTING_SOURCE=museum photo--> @DerekSikes what should I do with http://arctos.database.museum/SpecimenResults.cfm?collecting_source=museum%20photo
• COLLECTING_SOURCE=unknown-->no change
• COLLECTING_SOURCE=wild caught-->COLLECTING_SOURCE in situ
• COLLECTING_SOURCE=captive-->COLLECTING_SOURCE ex situ + add attribute establishment means=captive
• attribute experimental value='yes'--->attribute establishment means value='experimental'
• attribute experimental value='no'--->DELETE

Cleanup

• delete attribute 'experimental'
• delete COLLECTING_SOURCE=field photo
• delete COLLECTING_SOURCE=museum photo
• delete COLLECTING_SOURCE=wild caught
• delete COLLECTING_SOURCE=captive

check code for places collecting_source is used, fix as necessary

[campmlc]

Is the captive and experimental data going into an attribute called
"establishment"? Can we come up with a better name? That does not make any
sense to me at all. What about an attribute called "source"? That at least
has some recognizable meaning.

On Wed, Mar 20, 2019 at 1:36 PM dustymc [e-mail] wrote:

I think this is ready to go, except museum photo for which I'll do
something random and throw that in some remarks field somewhere if it's
still hanging around. Here's the revised plan:
Data

http://arctos.database.museum/info/ctDocumentation.cfm?table=CTCOLLECTING_SOURCE

gets two new terms:

• in situ==>inside, on site, or in the natural position in time and
  space
• ex situ==>outside, off site, or away from the natural position in
  time and space

New attribute "establishment means" definition "Refinement of
collecting_source." linked to new code table ctestablishment_means with
values

• captive --->Specimen was taken from captivity.
• experimental-->organism and its host do not represent a natural
  occurrence

Migration Path

• COLLECTING_SOURCE=field photo-->COLLECTING_SOURCE in situ
• COLLECTING_SOURCE=museum photo--> @DerekSikes
  https://github.com/DerekSikes what should I do with
  http://arctos.database.museum/SpecimenResults.cfm?collecting_source=museum%20photo
• COLLECTING_SOURCE=unknown-->no change
• COLLECTING_SOURCE=wild caught-->COLLECTING_SOURCE in situ
• COLLECTING_SOURCE=captive-->COLLECTING_SOURCE ex situ + add
  attribute establishment means=captive
• attribute experimental value='yes'--->attribute establishment means
  value='experimental'
• attribute experimental value='no'--->DELETE

Cleanup

• delete attribute 'experimental'
• delete COLLECTING_SOURCE=field photo
• delete COLLECTING_SOURCE=museum photo
• delete COLLECTING_SOURCE=wild caught
• delete COLLECTING_SOURCE=captive

check code for places collecting_source is used, fix as necessary

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-474948714,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AOH0hHCo6wDxKTHeJRztqkswyrJduerkks5vYnGDgaJpZM4bH33n
.

[dustymc]

https://terms.tdwg.org/wiki/dwc:establishmentMeans

[campmlc]

Ah, OK then. Sounds very plant centered, but as long as it has some meaning
to others outside of Arctoslandia I can live with it.

On Wed, Mar 20, 2019 at 11:31 PM dustymc [e-mail] wrote:

https://terms.tdwg.org/wiki/dwc:establishmentMeans

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-475103356,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AOH0hHpuvEsVqhVCIoMy0hhsiqhyY6Htks5vYv0EgaJpZM4bH33n
.

[dustymc]

plant centered

We don't have to use the terminology or vocabulary, it's just the closest sorta-similar thing I was familiar with. I think we'll map COLLECTING_SOURCE to DWC:establishmentMeans so maybe that'll just be confusing.

This can change back to next task when we have an attribute name and existing COLLECTING_SOURCE values mapped to new attribute values.

[Jegelewicz]

I like "establishment means" for the attribute. Is it possible that we could send a concatenation of COLLECTING_SOURCE and "establishment means" as dwc:establishment means? or if COLLECTING_SOURCE is ex-situ, send "establishment means" value instead?

[dustymc]

We can do WHATEVER for DWC, but that might break my hypothetical auto-generated "map of Whatever County [non]natives" (for which I can just use collecting_source). I suppose I could-maybe-should check for CONTAINS instead of IS if all I care about is the generalization.

Mapping and such aside, I think the clarification in the attribute is more critical when COLLECTING_SOURCE is ex-situ - is this a ringed seal in the San Diego Zoo (in which case I probably want to ignore it for most any mapping purposes) or a salamander wandering around the parking lot (in which case I definitely want it for my county-scale checklist, but not for my meter-scale habitat analysis).

[Jegelewicz]

We can do WHATEVER for DWC, but that might break my hypothetical auto-generated "map of Whatever County [non]natives" (for which I can just use collecting_source). I suppose I could-maybe-should check for CONTAINS instead of IS if all I care about is the generalization.

Couldn't you just use "establishment means" once we get that set up for your mapping?

I think the clarification in the attribute is more critical when COLLECTING_SOURCE is ex-situ - is this a ringed seal in the San Diego Zoo (in which case I probably want to ignore it for most any mapping purposes) or a salamander wandering around the parking lot (in which case I definitely want it for my county-scale checklist, but not for my meter-scale habitat analysis).

A parking lot isn't necessarily ex-situ and it probably will depend on who is cataloging how it gets recorded....but I guess we could learn to deal with that stuff if and when it arises?

[Jegelewicz]

I think the clarification in the attribute is more critical when COLLECTING_SOURCE is ex-situ

This is making me re-think. If you collect a eucalyptus in California, it might be hard to decide if in-situ or ex-situ is appropriate. However, in either case, an establishment means of introduced seems appropriate. Maybe we are trying to simplify something that isn't simple and perhaps we just need COLLECTING_SOURCE to be more robust....

[dustymc]

just use "establishment means"

We won't have that (yet!?) for most Occurrences; it's a refinement, not a primary concept.

parking lot isn't necessarily ex-situ

In the current model, that's a determination made by "us," and it's probably based on how maps worked 20 years ago and how the collector/cataloger sees the world and .... In what we're proposing, given useful data, the researcher can make that determination depending on what they're trying to get at.

who is cataloging

GIGO...

If we want to support meter-scale habitat analysis then we have to record meter-scale data and fine-grained metadata. This should make the metadata aspect of that possible.

hard to decide if in-situ or ex-situ is appropriate

Yup - so I might just ignore it in that case, or you might choose to enter "unknown."

establishment means of introduced

That's probably sorta useful, depending on what question I'm answering. An establishment means of introduced with a supporting method (esp. if it's a reference to a publication) and agent might let me get at what you ACTUALLY mean at a resolution useful for whatever I'm trying to do.

trying to simplify something that isn't simple

This will provide both: "simple" data that's useful for most things at normal/traditional scales (and that doesn't add anything to our workload), and the opportunity (not requirement) to also include as much supporting complexity as you have. Your eucalyptus might be "introduced" (because publicationA defines that term in some way that's relevant to this particular specimen) and "cultivated" (publicationB) and "naturalised" (publicationC) and ....

[Jegelewicz]

This will provide both: "simple" data that's useful for most things at normal/traditional scales (and that doesn't add anything to our workload), and the opportunity (not requirement) to also include as much supporting complexity as you have. Your eucalyptus might be "introduced" (because publicationA defines that term in some way that's relevant to this particular specimen) and "cultivated" (publicationB) and "naturalised" (publicationC) and ....

Agree, so what do we need to proceed?

an attribute name and existing COLLECTING_SOURCE values mapped to new attribute values.

New attribute "establishment means" definition "Refinement of collecting_source." linked to new code table ctestablishment_means with values

• captive --->Specimen was taken from captivity.
• experimental-->organism and its host do not represent a natural occurrence

Migration Path

• COLLECTING_SOURCE=field photo-->COLLECTING_SOURCE in situ
• COLLECTING_SOURCE=museum photo--> @DerekSikes what should I do with http://arctos.database.museum/SpecimenResults.cfm?collecting_source=museum%20photo
• COLLECTING_SOURCE=unknown-->no change
• COLLECTING_SOURCE=wild caught-->COLLECTING_SOURCE in situ
• COLLECTING_SOURCE=captive-->COLLECTING_SOURCE ex situ + add attribute establishment means=captive
• attribute experimental value='yes'--->attribute establishment means value='experimental'
• attribute experimental value='no'--->DELETE

Cleanup

• delete attribute 'experimental'
• delete COLLECTING_SOURCE=field photo
• delete COLLECTING_SOURCE=museum photo
• delete COLLECTING_SOURCE=wild caught
• delete COLLECTING_SOURCE=captive

I'm good with that.

[dustymc]

need to proceed

I think just the terminology - @campmlc can you suggest something or flip this to next task?

[campmlc]

These are somewhat overlapping - here's a first attempt at some
clarification based on what I've seen in the Rausch records. Maybe this
needs to be split further?

• captive ---> Specimen from a captive organism or population; may or
  may not have been bred in captivity; not a natural occurrence record;
• captive-bred--> Specimen from a captive organism or population; bred
  in captivity; not a natural occurrence record
• experimental-->organism was experimentally manipulated, applies also
  to host/parasite or host/pathogen life cycles maintained in laboratory
  cultures and for experimental infections; may nor may not have been
  captive-bred; not a natural occurrence record

For example, several mice could have been caught on St. Lawrence Island and
brought in the lab to start an experimental colony. Some might have been
immediately infected with a strain of Echinococcus taken from a St.
Lawrence Island fox. These would be wild-caught, but experimental. Others
would have been bred several generations and these also then infected with
Echinococcus that had been passed through several generations of mice/fox.
These would be captive-bred, experimental. A fox that was caught in the
wild and then kept for a year or two prior to being used in experiment
would be captive, experimental, but not captive-bred . . . ?

On Thu, Mar 21, 2019 at 12:44 PM dustymc [e-mail] wrote:

need to proceed

I think just the terminology - @campmlc https://github.com/campmlc can
you suggest something or flip this to next task?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-475309786,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AOH0hDqYCIDvBUa0p-xk6DgpIL41HdIBks5vY7brgaJpZM4bH33n
.

[dustymc]

Are there existing data you'd want to migrate differently than I've laid out?

Like all Attributes, this is data-driven - adding more values later isn't a problem.

I'm not sure it's ideal, but I don't see why this wouldn't work for that. If there's sufficient data the capture would be better as an event (and it's in situ at that point), then subsequent samplings/killing/whatever could be more events. If you don't have those data, this might work as 'lightweight events."

[ccicero]

Here's a suggestion based on our issues discussion 4/4/2019:

'Wild-caught' doesn't apply broadly, and is essentially the same (or has the same intent) as specimen_event_type of 'collection.'

'Field photo' and 'museum photo' are not helpful and really should be part of media. There are also some issues with those data. e.g., MSB Bird 43367 has a source of 'field photo' but it's a skeleton with tissue. ???? http://arctos.database.museum/guid/MSB:Bird:43367

What if we
1) add "captive" as another specimen_event_type ( could also extend to things like "cultivated"), with good definitions.
2) Clean up the data for 'field photo' and 'museum photo'
3) Get rid of 'collecting_source' altogether.
4) Use 'specimen_event_type' to determine what's mappable (e.g., 'collection' = yes, 'captive' = no)

https://arctos.database.museum/info/ctDocumentation.cfm?table=CTSPECIMEN_EVENT_TYPE

More specific details such as 'in situ' versus 'ex situ' could/should go into attributes for the specimen,

[ewommack]

Carla's suggestion makes sense to me. I don't think we'd loose anything for our collection if we got rid of collection source that we couldn't get from Specimen event type.

I think we should keep the collection method field though.

[dustymc]

Yea the photos thing definitely needs to go!

• You find a fossil specimen in the rock, and another similar specimen broken off a foot below the first. (They have different scientific value, one is only useful for broad-scale mapping, etc.) How does the proposal handle that?
• You catch a critter, huck it in a cage, periodically 'encounter' (eg, biopsy) it, eventually kill it. Another critter is encountered in nature and then killed. How do I distinguish them?

[Jegelewicz]

Here is what I would do:

Scenario #1

You catch a critter, huck it in a cage

Collection event (whole organism) (part condition=alive)

periodically 'encounter' (eg, biopsy) it

Each is a "Captive" event (blood, ear clip, etc.) with a separate catalog number (use same individual as to relate to above)

eventually kill it

"Captive" event (whole organism) with a separate catalog number (use same individual as to relate to above)

Scenario #2

Another critter is encountered in nature

"Collection" event (blood, ear clip, etc.) each event with a separate catalog number (use same individual as to relate to each other)

and then killed.

"Collection" event (whole organism) with a separate catalog number (use same individual as to relate to above)

How do I distinguish them?

Collection events are mappable, captive events are not

[Jegelewicz]

You find a fossil specimen in the rock, and another similar specimen broken off a foot below the first. (They have different scientific value, one is only useful for broad-scale mapping, etc.) How does the proposal handle that?

I don't know how often this happens. Why wouldn't they both be useful for mapping?

[dustymc]

The captivity thing works for me, as long as we're willing to assert opinions as entire event stacks and add 'cultured' and 'introduced' and 'invasive' and etc. as they come in. I'm concerned about finding mutually exclusive definitions for those, which is a distant-secondary problem (because they're secondary data) as this was proposed.

I don't know how often this happens. Why wouldn't they both be useful for mapping?

n=1, but I think float is a huge part of most paleo collections. For "fossils of whatever county" it doesn't matter at all. For doing fine-scale analyses, including fine-scale mapping, the one that has a KNOWN context is valuable and one without is not.

mappable

That's the biggie for me. In what I proposed, users can fine-tune that based on things like the precision of the data they're working with. In the current model and what's being proposed here it's a curatorial assertion, which means it's close to useless for anyone who sees the world at a different scale then whoever assigned the value.

[Nicole-Ridgwell-NMMNHS]

During the meeting we talked briefly about putting the in-situ/ex-situ data (or refinements thereof) in specimen attributes instead. I think I like that solution

[Nicole-Ridgwell-NMMNHS]

I don't know how often this happens. Why wouldn't they both be useful for mapping?

In this particular scenario, probably yes. Most of the time though, you don't really know which bed the float came out of unless you find the other parts of the specimen still in situ.

[dustymc]

in-situ/ex-situ data (or refinements thereof) in specimen attributes

I still think that's backwards.

With many legacy specimens, data is going to be light. "It's probably about where you'd expect to find it," "It's probably not about where you'd expect to find it," "we don't know." From @Jegelewicz eucalyptus example, something maybe-arbitrary happens there and under my proposal we can use the attributes to record "introduced" and link it to methods and people etc., "invasive" to other methods and people, etc. With the fine-scale data being in the forefront you have to determine that ahead of time, even if there's no universally-agreed answer, additional determinations drag along an entire locality stack (which seems like something many of us are trying to avoid), and you still don't have a great place for "establishmentmeans according to AGENT using METHOD."

useful for mapping

Again, this depends on scale. If you're recording localities at the scale of kilometers then I don't think this is so important, but if you're recording localities at the scale of cubic centimeters then the utility of the specimen depends on the layers in your map. (And that "15 feet up a formation" thing is in the primary spatial data where it belongs.)

And I think I want to go back on my "the captive thing seems OK" statement, although perhaps this isn't the place to do so. I think most users will assume "collect" is a final act - any samples from the same animal collected at a later date are suspect. "Collect" (and then maybe let go later) and "collect" (and put the skull in a museum) are indistinguishable in the above example. Perhaps a "captive" specimen_event_type would be useful for 'removed from nature, didn't kill."

[campmlc]

I agree with this: Perhaps a "captive" specimen_event_type would be useful
for 'removed from nature, didn't kill."

—

On Thu, Apr 4, 2019 at 5:13 PM dustymc [e-mail] wrote:

in-situ/ex-situ data (or refinements thereof) in specimen attributes

I still think that's backwards.

With many legacy specimens, data is going to be light. "It's probably
about where you'd expect to find it," "It's probably not about where you'd
expect to find it," "we don't know." From @Jegelewicz
https://github.com/Jegelewicz eucalyptus example, something
maybe-arbitrary happens there and under my proposal we can use the
attributes to record "introduced" and link it to methods and people etc.,
"invasive" to other methods and people, etc. With the fine-scale data being
in the forefront you have to determine that ahead of time, even if there's
no universally-agreed answer, additional determinations drag along an
entire locality stack (which seems like something many of us are trying to
avoid), and you still don't have a great place for "establishmentmeans
according to AGENT using METHOD."

useful for mapping

Again, this depends on scale. If you're recording localities at the scale
of kilometers then I don't think this is so important, but if you're
recording localities at the scale of cubic centimeters then the utility of
the specimen depends on the layers in your map. (And that "15 feet up a
formation" thing is in the primary spatial data where it belongs.)

And I think I want to go back on my "the captive thing seems OK"
statement, although perhaps this isn't the place to do so. I think most
users will assume "collect" is a final act - any samples from the same
animal collected at a later date are suspect. "Collect" (and then maybe let
go later) and "collect" (and put the skull in a museum) are
indistinguishable in the above example. Perhaps a "captive"
specimen_event_type would be useful for 'removed from nature, didn't kill."

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-480095739,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AOH0hMo-Fs4_PdGrfK9bq8iOWBS9H3W6ks5vdocsgaJpZM4bH33n
.

[dustymc]

Can I go ahead with https://github.com/ArctosDB/arctos/issues/1942#issuecomment-475308419 then?

[campmlc]

Yes, agree.

[ewommack]

The more I looked into it there appears to be a debate about the in situ/ex situ use in different disciplines.
I'm really not sure that choice is the best to replace the current collection sources with.

[dustymc]

@ewommack replacement suggestions?

[ewommack]

We had a discussion in the issues meeting of the working group, and I like Carla's suggestion (https://github.com/ArctosDB/arctos/issues/1942#issuecomment-480017229)

[Jegelewicz]

From @dustymc

The new suggestion does not address the problem. “Collected”==’killed, found dead, or removed from functional cultural, biological, ecological, or archeological context.’ An alligator can be “collected” from the Yukon River. Collection_source says “we don’t think it got there by itself.” The proposed attribute lets us debate the finer points of how it got there (introduced, invasive, etc.). “what's mappable (e.g., 'collection' = yes, 'captive' = no)” depends on things like the precision of the data; the original issue was about being able to do much more with the data, I think the new proposal is at best a wash letting us do the same things under a slightly different vocabulary.

[dustymc]

AWG

• add to collecting source as necessary

• wild caught-->wild

• still need the attribute for specifics?

[dustymc]

field photo

curators will deal with this. If they don't by (date to be picked at next AWG meeting) then we'll do this:

• change to wild [caught??]
• leave a remark

/remind me to add a date next thursday

[reminders[bot]]

@dustymc set a reminder for May 9th 2019

[ccicero]

Carla cleaned up MVZ records - changed all 'field photo' to 'wild caught' which may be changed to 'wild' (see previous comment).

[dustymc]

I'm dragging https://github.com/ArctosDB/arctos/issues/2068#issuecomment-489182217 and https://github.com/ArctosDB/arctos/issues/2068#issuecomment-489183841 back over here.

I really hope that we will all start to see our data as useful to more than a single audience.

That starts with recording data that broader audience can access as THEY want to, and not just as "we" thought they might want to.

wild

Here's a real use case. There are three (or 4 or more, depending on how you feel about chronospecies) flavors of bison in AK - 'native' wood bison which went extinct around 1900, introduced plains bison, and reintroduced (from CA) wood bison. The plains bison probably wouldn't make it without farming, so it might be justifiable to call them something like 'cultivated.' The recent wood bison folks might tell you the plains bison are both introduced and invasive. Some would object to the word 'reintroduced' because the wood bison are not quite the native stock. Etc.; it's complicated. I think most would agree that they're all "wild." Various users are likely to want various bit-n-pieces of them (or not) in various datasets for various purposes.

Given a frozen bison-bit, someone in the collection isn't likely to know more than "probably not from the zoo" and the best they can do is use 'wild' as collecting source.

Collecting source is the "[not] wild" component. I don't think anything about this would change that, just the vocabulary. Having a way for mammal-people to say "valid for most mapping purposes" and a different way for paleo-people to say "valid for most mapping purposes" and ... isn't ideal, but we can work with it. (And as @campmlc suggested maybe we can somehow synonymize terms in the interfaces.)

The attribute proposed in https://github.com/ArctosDB/arctos/issues/1942#issuecomment-466558248 provides a place to get at the finer details, should anyone eventually record them. (It's introduced because some will see it that way, it's native because others will see it that way, it's cultivated-er-something because F&G is chucking hay to them, etc., etc., etc.)

The attribute also provides a place to clarify (or argue...) if there's legitimate disagreement over 'wild' - maybe whoever's entering data doesn't consider plains bison in AK (camels in AU, rats sorta anywhere, etc., etc., etc.) to be 'wild' - the attribute can deal with that.

The attribute can also deal with fine-scale data: a paleo-bit 2 feet below where it probably came from, a parking-lot salamander that's a great county record but not so great for some kinds of habitat analysis, etc.

The workload for entering a "normal" specimen would not change - MAYBE it would be different vocabulary, but that's it. If someone shows up with fine-grained data (or opinions about why there are eucalypts in CA, or ...) then we have a place for it in the attribute.

I'm not sure that adds up to something that can be easily filtered before a search, but we should definitely find some way of conveying the complexity of the data to users. Filtering on collecting source will make most users happy, and maybe we can figure out how to colorcode maps by collecting source or something.

If that all works, I think this is the path forward:

• create a new attribute "establishment means," move 'experimental' data to it
• add 'exhumed' (and whatever else anyone wants, within reason) to http://arctos.database.museum/info/ctDocumentation.cfm?table=CTCOLLECTING_SOURCE
• clean up '... photo' values
• maybe flip 'wild caught' to 'wild'

What did I miss?

[campmlc]

I like this.
I also like the idea of "colorcode maps by collecting source" - but that
conflicts with Berkeley Mapper's current color coding. However, people are
confused by the current BM colors - it would be good to have a map with a
legend/key.

On Fri, May 3, 2019 at 5:58 PM dustymc [e-mail] wrote:

I'm dragging #2068 (comment)
https://github.com/ArctosDB/arctos/issues/2068#issuecomment-489182217
and #2068 (comment)
https://github.com/ArctosDB/arctos/issues/2068#issuecomment-489183841
back over here.

I really hope that we will all start to see our data as useful to more
than a single audience.

That starts with recording data that broader audience can access as THEY
want to, and not just as "we" thought they might want to.

wild

Here's a real use case. There are three (or 4 or more, depending on how
you feel about chronospecies) flavors of bison in AK - 'native' wood bison
which went extinct around 1900, introduced plains bison, and reintroduced
(from CA) wood bison. The plains bison probably wouldn't make it without
farming, so it might be justifiable to call them something like
'cultivated.' The recent wood bison folks might tell you the plains bison
are both introduced and invasive. Some would object to the word
'reintroduced' because the wood bison are not quite the native stock. Etc.;
it's complicated. I think most would agree that they're all "wild." Various
users are likely to want various bit-n-pieces of them (or not) in various
datasets for various purposes.

Given a frozen bison-bit, someone in the collection isn't likely to know
more than "probably not from the zoo" and the best they can do is use
'wild' as collecting source.

Collecting source is the "[not] wild" component. I don't think anything
about this would change that, just the vocabulary. Having a way for
mammal-people to say "valid for most mapping purposes" and a different way
for paleo-people to say "valid for most mapping purposes" and ... isn't
ideal, but we can work with it. (And as @campmlc
https://github.com/campmlc suggested maybe we can somehow synonymize
terms in the interfaces.)

The attribute proposed in #1942 (comment)
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-466558248
provides a place to get at the finer details, should anyone eventually
record them. (It's introduced because some will see it that way, it's
native because others will see it that way, it's cultivated-er-something
because F&G is chucking hay to them, etc., etc., etc.)

The attribute also provides a place to clarify (or argue...) if there's
legitimate disagreement over 'wild' - maybe whoever's entering data doesn't
consider plains bison in AK (camels in AU, rats sorta anywhere, etc., etc.,
etc.) to be 'wild' - the attribute can deal with that.

The attribute can also deal with fine-scale data: a paleo-bit 2 feet below
where it probably came from, a parking-lot salamander that's a great county
record but not so great for some kinds of habitat analysis, etc.

The workload for entering a "normal" specimen would not change - MAYBE it
would be different vocabulary, but that's it. If someone shows up with
fine-grained data (or opinions about why there are eucalypts in CA, or ...)
then we have a place for it in the attribute.

I'm not sure that adds up to something that can be easily filtered before
a search, but we should definitely find some way of conveying the
complexity of the data to users. Filtering on collecting source will make
most users happy, and maybe we can figure out how to colorcode maps by
collecting source or something.

If that all works, I think this is the path forward:

• create a new attribute "establishment means," move 'experimental'
  data to it
• add 'exhumed' (and whatever else anyone wants, within reason) to
  http://arctos.database.museum/info/ctDocumentation.cfm?table=CTCOLLECTING_SOURCE
• clean up '... photo' values
• maybe flip 'wild caught' to 'wild'

What did I miss?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/ArctosDB/arctos/issues/1942#issuecomment-489273585,
or mute the thread
https://github.com/notifications/unsubscribe-auth/ADQ7JBBK2P5GBTZQGU22ALTPTTGQPANCNFSM4GY7PXTQ
.

[Jegelewicz]

If that all works, I think this is the path forward:

create a new attribute "establishment means," move 'experimental' data to it
add 'exhumed' (and whatever else anyone wants, within reason) to http://arctos.database.museum/info/ctDocumentation.cfm?table=CTCOLLECTING_SOURCE
clean up '... photo' values
maybe flip 'wild caught' to 'wild'

Sounds good to me. Now I need to go back to the paleo people and find out what they want. I think they were fans of in situ/ex situ...

[dustymc]

Specimen_event.collecting_source is NULLable, and the code table contains "unknown." Is there a need for "we implicitly don't know" and "we explicitly don't know" or something, or can we merge those one way or the other?

[reminders[bot]]

:wave: @dustymc, add a date

[dustymc]

AWG2019_05_09: Declined to set firm deadline.

Here are collections using 'field photo'.

  guid_prefix, count(*)
from 
  collection,
  cataloged_item,
  specimen_event
where
  collection.collection_id=cataloged_item.collection_id and 
  cataloged_item.collection_object_id=specimen_event.collection_object_id and
  specimen_event.COLLECTING_SOURCE='field photo'
group by
 12    guid_prefix;

GUID_PREFIX                            COUNT(*)
------------------------------------------------------------ ----------
UAMObs:Ento                              78
DMNS:Para                                 1
UAM:Ento                                  1
UCM:Obs                                  35
NMU:Mamm                                  1
UAMObs:Mamm                              27
UAMObs:Fish                               8

[Jegelewicz]

@DerekSikes @acdoll @keg34 Can you check on the above and correct? Not sure who else to ping.

[dustymc]

Dropping critical tag; this is not breaking existing functionality.

This could add a great deal of functionality to Arctos - let's somehow not let it die.

[campmlc]

Working with parasitologists at HWML. They have data fields for "experimental" vs "natural". Perhaps we change "experimental, yes/no" as an attribute to a new attribute "collection source" with possible values: "natural, experimental, captive, collected in situ, collected ex situ . . . " , and a given record could have more than one of these attribues. Just throwing this out. The attribute would be associated with the specimen, not the specimen event. Again, just thinking fast, this makes sense because the collection event should properly be shared between related items such as host/parasite, but the collection source may not be shared, e.g. the host might be wild caught = field collected, but the parasite could potentially be experimental . . .

[dustymc]

@campmlc is that different than https://github.com/ArctosDB/arctos/issues/1942#issuecomment-473356057 ?

[campmlc]

Only in the sense of adding more terms that would capture data more relevant to parasite collections, including experimental/natural, and confirming that the attribute model of collection source would work for this situation.

[dustymc]

Yes, and yes albeit imperfectly.

https://github.com/ArctosDB/arctos/issues/1942#issuecomment-466558248

[campmlc]

Should experimental be an option for establishment means? The worm/virus/plant did not get into this this place by itself. It was put there on purpose.

[Jegelewicz]

See also https://github.com/tdwg/dwc/issues/235