ArctosDB / arctos

Arctos is a museum collections management system
https://arctos.database.museum
61 stars 13 forks source link

Place is not working for OGL #8269

Closed happiah-madson closed 2 weeks ago

happiah-madson commented 2 weeks ago

Help us understand your request (check below):

Describe what you're trying to do OGL needs to be able to assert both marine and terrestrial place fields on our records. Why do we need to be able to assert both? 1) Permitting is done by terrestrial things. (The Atlantic Ocean does not issue a permit. Massachusetts Department of Marine Fisheries issues a permit for collecting in the Atlantic Ocean that is on MA's coast.) 2) We are a marine collection and it is critical that we are able to query and understand our records using marine bodies of water.

Because of the second constraint, all of our higher geography is all ocean-related. And yet, this is what our "continent/ocean" data looks like when I look at catalog record summary:

Screen Shot 2024-11-05 at 10 09 45 AM

I honestly have no idea what this is pulling from, but the results certainly don't match the definition: "The continent or ocean within which a locality is found. This is the highest geographic category." because the Gulf of Maine is not an ocean. It also asserts that we have have 963 records in the Pacific Ocean. In our old database, we had over 2,000 locations in the Pacific Ocean. So even our attempt to have things "standard" by putting Ocean in higher geography has not rendered our records useful.

If there are no changes made to geography we have two options:

  1. Continue to put ocean into higher geography.

Here are the consequences: all terrestrial information (continent, country, state, city, island, island-group) must be put into specific locality, along with any other marine information, such as the specific waterbody (such as Gulf of Maine). This requires us to put something into our record bulkload template to tell people to put all of this information into specific locality, which, to my mind is the same things as treating it as remarks. If we export data we then have to spend time pulling the standardized information in specific locality (such as continent, country, state, city, island, island group, ocean, specific waterbody) out of a mess of a field before it is useful.

  1. Convert all our higher geographies to GADUM country, state, where applicable. For records in open waters, continue to use ocean.

Here are the consequences: some terrestrial information (country, state) is standardized for records. However, continent, island, island-group, ocean, specific waterbody would need to be put in specific locality. This still requires us to put something into our record bulkload template to tell people to put all of this information into specific locality, and if we export data we then have to spend time pulling the standardized information in specific locality out of a mess of a field before it is useful.

There is a difference between Collection-Asserted information and cool background stuff that improves findability of records and has shape files for people who are looking at our data. I need to ASSERT and be able to confirm that things are properly ASSERTED on my records in order to ensure that we can find things in ways we expect to.

It is for this reason that @genevieve-anderegg @sharpphyl Rosie and I feel so strongly about waterbody! Let me assert the things that I need to assert! I'm happy to meet to talk about this, but Arctos's marine collections are in agreement about needing this field or other ways to assert important marine things about our collections. It's not productive for us to have any more geography meetings without @mkoo and @dustymc so if you need a meeting, please send a poll for setting up a meeting.

dustymc commented 2 weeks ago

I think I see two functional needs in the above:

  1. You want to assert the normal '{event_typed} there' place
  2. You additionally want to assert - well, I'm lost, I'll wander around below.

(1) I believe is generally collected from some sort of watery place (such as https://arctos.database.museum/place.cfm?action=detail&geog_auth_rec_id=10016562). This all seems straightforward, although perhaps we'd need to find more authority data (eg something which contains the parent of the example I found while trying to match your example!) or rebuild the geography model (our continent_ocean field ain't helping clarify anything...) or whatever.

(2) I think you're saying you want to mention eg https://arctos.database.museum/place.cfm?action=detail&geog_auth_rec_id=1018 even though that's not where anything happened, and the patch of rocks-n-dirt encompassed by that polygon definitely is not capable of issuing a permit, hence my complete bafflement. I suspect you simply need to record that https://arctos.database.museum/agent/21350125 issued a permit, but that's not (quite) what I'm hearing and I don't understand why or how that involves geography.

happiah-madson commented 2 weeks ago

Okay, here are my thoughts:

  1. You want to assert the normal '{event_typed} there' place

Yes, I absolutely do want to do this.

You additionally want to assert [...] a watery place. [...] we'd need to find more authority data (eg something which contains the parent of the example I found while trying to match your example!)

Yes, I want to be able to always and systematically assert a watery place. I'm not sure that more authority data (or a more authoritative source) exists and the stress of trying to create and curate an authority is why I believe we were decided to not try and link different watery places with shape files.

I think you're saying you want to mention eg https://arctos.database.museum/place.cfm?action=detail&geog_auth_rec_id=1018 even though that's not where anything happened, and the patch of rocks-n-dirt encompassed by that polygon definitely is not capable of issuing a permit, hence my complete bafflement.

Yes, I do also want to be able to assert the closest terrestrial information, when applicable. I do think that you are right that I need to record that https://arctos.database.museum/agent/21350125 issued a permit as a separate action, but the waters outside of the MA polygon in https://arctos.database.museum/place.cfm?action=detail&geog_auth_rec_id=1018 are still part of MA! If I do a search for MA, I need records from MA waters to be included.

dustymc commented 2 weeks ago

I've clearly failed at extracting functional requirements and don't quite know where to go from there, but that still seems like a necessary prerequisite to any real progress.

I have left some possibly-relevant comments in https://github.com/orgs/ArctosDB/discussions/7666.

dustymc commented 2 weeks ago

Very high overview: I think we've agreed (here: https://docs.google.com/document/d/1aEzwbnDObmnjKkkyM9BVMqWhKACPloZenfc9kpMHxHo/edit?tab=t.0) to move (not replicate/confuse/denormalize - yay!) all watery-stuff (perhaps as a first step) to locality attributes.

Is there anything else to resolve with this?

happiah-madson commented 2 weeks ago

I don't think so?

genevieve-anderegg commented 2 weeks ago

Maybe some UI improvements on how this would be displayed on the catalog record page, but that is a separate request for later once we get the data moved around! Thank you so much Dusty!!!

dustymc commented 2 weeks ago

Tentatively closing then. (And don't thank me, I'm just pushing buttons and absolutely never would have got where I think we're going - which is starting to look like a good place to be - by myself, thanks to all of you!)

I've added some ramblings to https://github.com/orgs/ArctosDB/discussions/7666

Yes of course UI changes, possibly mapping things to FLAT (for easier "local" access) and/or DWC (https://github.com/ArctosDB/arctos/issues/7348), etc., etc. would all come along with this (somehow, possibly ripping the bandaid off would mean easier UI than the evolutionary process I think we've begun).