[Tagging] multiple schools on one plot
steveaOSM at softworkers.com
Mon Apr 5 22:30:36 UTC 2021
I don't know if it is TOO late (to change the status quo), this is OSM, after all, where we have plastic tagging and processes in place like the ability to formally propose new tags, restructure present tags, etc. However, it is what I might call "very" late, as there are thousands of legacy (existing, really) amenity=school tags on both polygons and nodes that would have to be "unsnarled" — whether in a query-by-query approach (which I agree seems distasteful long-term as it is "snarly" each time we have to do it now) or in a "global" approach where we think about it now and figure out a smart method of how to disambiguate, then propose, then redact what we decide is legacy (if we do that).
Either way (or another), this takes some thought, collaboration, agreement, work and (wiki-)writing. I am disheartened that there has been so much misunderstanding (which I partially attribute to how "usage happened" but the wiki didn't quite keep up by properly reflecting said usage) and even rancor on the topic. However, I believe OSM can forge ahead and find an amicable long-term solution. I believe I have stated (repeatedly!) my reasonings and because my points have been made, I'm not sure I have more to contribute. Are there now 65 (!) messages in this single thread? Whew!
Thanks, Minh, for splitting off a new thread here that co-mingles the important consideration of mixing landuse=religion with the possibility of a more-well-defined landuse=education (or school) to complement or replace amenity=school.
> On Apr 5, 2021, at 1:56 PM, Minh Nguyen <minh at nguyen.cincinnati.oh.us> wrote:
> This is the cruxt of the issue: we know that amenity=school already has these two definitions and it's difficult to distinguish between the two in queries. Is there an opportunity to split it up into two tags? Or is it too late to change the status quo to benefit use cases we didn't foresee over a decade ago when amenity=school started being used this way? If amenity=school had come into popular use more recently, would renderer developers have declined to support it because of its varied scope, as with amenity=research_institute? 
>  https://github.com/gravitystorm/openstreetmap-carto/issues/2077
> Vào lúc 08:00 2021-04-05, stevea đã viết:
>> On Apr 5, 2021, at 7:43 AM, Jmapb <jmapb at gmx.com> wrote:
>>> On 4/2/2021 7:07 AM, Mateusz Konieczny via Tagging wrote:
>>>> It avoid using fake amenity=school
>>>> If there are 2 schools in one complex, then there should be 2
>>>> objects there - not 1, not 3 and not 23.
>>> This is key IMO. Ideally, I should be able to answer the question "How
>>> many schools are in this region?" by querying Overpass for
>>> amenity=school. If our recommended tagging scheme yields the answer of 3
>>> when only 2 actual schools exist, then we need better a better scheme.
>> No, not so, at least with such an overly-simplistic (and hence incorrect) query, since it makes specious assumptions: as amenity=school applied to a (multi)polygon can mean EITHER "school" OR (importantly) "school grounds," (especially with multiple schools on such a single site, e.g. if there are multiple nodes tagged amenity=school inside of the (multi)polygon also so tagged), one making such a query would need to determine whether or not this convention is extant in the query results.
>> Sometimes queries can be, and therefore are, quite straightforward, yielding correct results from an easy query. In this case, because of our conventions (as I thanked Minh for stating in my last post here), such a query is not quite as straightforward. This isn't anyone's fault in particular, simply a result of OSM having established a convention with applying the amenity=school tag to (multi)polygons having a distinct (and slightly different) meaning separate from applying amenity=school to a node.
More information about the Tagging