Proposal talk:Find A Grave cemetery and grave IDs

Licensing

What is the license of "find a grave"? Seems to be a proprietary database with "Copyright © 2021 Find a Grave" at the bottom. Mateusz Konieczny (talk) 07:34, 12 March 2021 (UTC)

I’ve seen references to Find A Grave entries elsewhere en masse, but I’ll double-check OSM’s policy as well as contact Ancestry.com about this just to make sure. —Sterling (talk) 15:21, 12 March 2021 (UTC)

I would personally prefer for OSM to only include external links to first-party content (created by the operator of a particular point of interest) and open-licensed content. Not even because of legal issues, but because cooperating with a proprietary, third-party platform doesn't really match the spirit of promoting open data. As far as I know, though, there's no standing policy on external links, so that's a matter of opinion. Feels a bit wrong to have the work of unpaid volunteers owned by one company. :( --Tordanik 18:29, 12 March 2021 (UTC)

Turns out the purely factual information of a person’s name, birth/death dates, cemetery location, etc. are uncopyrightable! The only thing that would be copyrightable (and would in fact remain in the hands of the volunteer) are the specific wordings of the descriptions of each cemetery, grave and person, as well as the images—all of which remain property of the volunteers. I cross-referenced several resources as well as contacted a site admin, but this article by the Legal Genealogist sums it up nicely. But yes, I agree wholly with your philosophy, and I doubt Find A Grave would be as widely adopted if their Terms of Service were any less permissive. —Sterling (talk) 19:12, 12 March 2021 (UTC)

So they are in jurisdiction without https://en.wikipedia.org/wiki/Database_right ? Are they supporting exporting of their database or blocking any attempt to copy it? Mateusz Konieczny (talk) 19:18, 12 March 2021 (UTC)

Correct; database rights aren’t a thing in the United States. As for exporting or the like: they don’t have a public API (that I know of), but high-speed scrapers have existed for years on end with no contest. —Sterling (talk) 23:13, 12 March 2021 (UTC)

Even if scraping the database is possible, morally I would think OSM as a community would wish to respect databases pulled together by others. "Do unto others as you wish them to do unto you". Historically, it has been believed to be important that OSM is seen to be above reproach in this matter: a single court case could bankrupt the foundation. SK53 (talk) 21:29, 25 March 2021 (UTC)

Also OSM's data is not held in the US, it is according to the laws of England & Wales which (at present) both allow database rights & allow copyright on 'sweat-of-brow' data collections. I think describing this as "Resolved" is inaccurate. SK53 (talk) 22:06, 26 March 2021 (UTC)

I have removed the resolved tag, as I cannot see how you can scrape data from Ancestry sites & also comply with items 1 & 2 in their Terms of Service. OSM is not a platform for casuistry w.r.t. to database rights. SK53 (talk) 22:13, 26 March 2021 (UTC)

Noone proposed importing Find a Grave database into OSM Mateusz Konieczny (talk) 23:23, 26 March 2021 (UTC)

Data is not being scraped. This proposal is merely a reference to external information, similar in function to contact:facebook=* or wikipedia=*. —Sterling (talk) 17:22, 29 March 2021 (UTC)

navigate linked data via Wikidata?

There will be debate whether an external id belongs in OSM or not. One choice is to delegate the id-broking to Wikidata. So, would a better alternative be to include a person's wikidata id on the grave then a program can retrieve the find-a-grave id (or Library of Congress id, or Hansard id, etc., etc.) from them rather than storing it on the OSM object? Jnicho02 (talk) 07:43, 12 March 2021 (UTC)

My only concern was the extra work creating Wikidata entries that would be required on part of those entering said IDs, since (at least in my case in rural West Virginia, United States) the vast majority of these cemeteries are only found on Find A Grave. Otherwise, I wholly agree that simplifying everything into an existing scheme would be far more intuitive. Also, refer to Jnicho02’s comment below about its state as a “well established” and “de facto standard reference”. —Sterling (talk) 15:21, 12 March 2021 (UTC)

I was personally thinking of making a search box for users to find and link Wikidata items in iD. Also editing Wikidata iteslf in iD could also be an option. The Wikidata community themselves have expressed great interest in being able to edit Wikidata from a map. I am also strongly opposed to adding duplicate data between Wikidata and OSM. Wikidata should be preferred where-possible since it is CC0 and has a better relational database model to begin with. --Lectrician1 (talk) 18:47, 22 March 2021 (UTC)

ref:

Resolved: Implemented as ref:find_a_grave=*.

This being an external reference, I would =:

• use Key:ref to give ref:find-a-grave and ref:find-a-grave:cemetery
• state that Find-A-Grave is international and therefore does not need an additional country code, e.g. ref:US:find-a-grave
• state that the reference is not verifiable on-the-ground (gravestones don't have Find-A-Grave ids carved on them)
• state that Find-A-Grave is so well established that it is effectively a de facto standard reference - Jnicho02 (talk) 08:14, 12 March 2021 (UTC)

These should build a stronger case for the proposal; I’ll implement these changes combined with that of Marc’s comment below. —Sterling (talk) 15:21, 12 March 2021 (UTC)

don't duplicate main tag into key again

I am not convinced that we need to repeat the name of the main attribute in the key, cemetery or grave is already given with the main attribute. Therefore ref:findagrave should be enought --Marc marc (talk) 10:13, 12 March 2021 (UTC)

That was my initial idea, but I didn’t know which was best for statistics and sorting. I’ll probably go with this, then. —Sterling (talk) 15:21, 12 March 2021 (UTC)

Is it possible that something tagged as cemetery in OSM is treated as grave in find a grave database? Is there both grave id 1000 and cemetery id 1000 or is there a single object with id 1000 in their database? Mateusz Konieczny (talk) 16:20, 19 March 2021 (UTC)

Grave and cemetery IDs are separate. Cemetery #123 would be located under https://www.findagrave.com/cemetery/123, whilst grave (“memorial”) #123 would be located under https://www.findagrave.com/memorial/123. —Sterling (talk) 01:47, 24 March 2021 (UTC)

So separate keys may be needed for cases where something is cemetery in OSM and grave in Find A Grave - or reverse. Note that "Find a grave classifies it as a grave, not a cemetery and it breaks linking" is not a valid reason to retag cemetery as a grave in OSM. Mateusz Konieczny (talk) 11:38, 24 March 2021 (UTC)

In what situation would this be possible? In either case, prefixing the value of the tag might help in this hypothetical situation, though that just adds another layer of complexity. —Sterling (talk) 16:41, 24 March 2021 (UTC)

Small cemetery classified as a grave on Find A Grave, or grave (possibly monumental one) missclassified as a cemetery on Find A Grave Mateusz Konieczny (talk) 18:28, 24 March 2021 (UTC)

Cemeteries must contain a grave, and graves must be contained within a cemetery. —Sterling (talk) 02:18, 25 March 2021 (UTC)

(1) even if true then Find A Grave may be simply wrong. (2) Newly opened cemetery may be without burials, there are cases of symbolic cemeteries or where graves were moved (3) there are many cases of standalone graves without being in a cemetery Mateusz Konieczny (talk) 19:20, 29 March 2021 (UTC)

small cemeteries : Notability for wikidata

I submitted the question on wikidata wiki Marc marc (talk) 21:56, 19 March 2021 (UTC)

extract from the wikidata wiki answer : Cemeteries do not require (...) any external identifier (...) to be notable for Wikidata (...) If a plot of land is verifiably recorded as a current or former cemetery in any credible print or electronic source, then it likely warrants a Wikidata item.

It is therefore not correct to use the argument of the notability of wikidata to justify the need for ref:find_a_grave on cemeteries, I have not yet asked the question for individual graves Marc marc (talk) 09:30, 20 March 2021 (UTC)

I agree, the argument is flawed and indeed there are over 30k graveyards and 6000 individual tombs in Wikidata at the moment and there is no reason why more cannot be added. Also Wikidata already has a property for an individual grave and for a cemetery which are modelled separately. Also importantly, there are already over 8000 cemeteries and over 98k items are linked to gravestones (even though many of them are actually linking people to their tombstone and are not items about the tomb itself), so this proposal would only create duplication. --Hannes Röst (talk) 18:57, 20 March 2021 (UTC)

I must have not read deeply enough into the standards, then; I’m not too familiar with Wikidata. The only issue then is the requirement of a “credible print or electronic source”, for which countless rural graveyards have none. If Find A Grave itself—a tertiary source—is fair game for citation, then I can’t help but somewhat agree that consolidating to Wikidata would be the best course of action. Let me know what you find out. —Sterling (talk) 01:47, 24 March 2021 (UTC)

"Find a Grave" is a registered trademark

I'm not sure if this is an issue, but I'm not pleased about tagging using a registered trademark as the keyword (per the Ancestry.com T&C). Related to this, I don't think it's wise to tag using a proprietary ID number. I would think it preferable to tag the name, birth date, and death date then use the bounding region data to fill in the cemetery location for the Find a Grave search. --JOlshefsky (talk) 11:11, 24 March 2021 (UTC)

"I'm not pleased about tagging using a registered trademark as the keyword" - why it would be problem if it is used to refer to that specific database? Why you link to ancestry.com TC? Is it related in some way to find a grave? Is there any reason to suspect that OpenStreetMap would be even influenced by this rules? I am not a lawyer, but I am betting that this conditions are at least in part void and unenforceable anyway. Mateusz Konieczny (talk) 11:36, 24 March 2021 (UTC)

the trademark isn't a issue, social networks are brands, that doesn't prevent to have keys with their name to make a link (which doesn't mean that this link is interesting for osm, it's another issue) Marc marc (talk) 20:38, 25 March 2021 (UTC)

The term “Find A Grave” would only be used in the same way “Facebook” is for contact:facebook=*. —Sterling (talk) 17:25, 29 March 2021 (UTC)