[Tagging] Layers (was Eruvs etc.)

[stevea]

On Sep 6, 2022, at 3:47 AM, Frederik Ramm <frederik at remote.org> wrote:
> Hi,
> On 05.09.22 22:24, Brian M. Sperlongano wrote:
>> So if you don't want to see religious boundaries when mapping trees in Karlsruhe, you don't have to - they're separate objects in separate layers that can't be comingled.

OK, here we see some of the wider aspects of a spectrum:  on the one hand, some say “come one, come all…put your data into OSM, whether it is natural, architectural, religious, part of the urban fabric, part of the urban-PLANNING fabric (doesn’t even yet exist, but it is proposed, or under construction…) and has that “wide, welcoming” aspect I mention.  While, on the other hand, some say “OSM is not a garbage dump for whatever you think is both in the real world and geosnatially-specific enough that it can and should be added to a shared fabric of reality in the guise of data we can choose to project onto screens as renderings or components of a route to get here from here to there (on a bicycle, on a subway, on a nature-hike…).

Mixed into that is the concept of “layers,” like whether we might CHOOSE to see “trees, nature, hilly contour, grass-in-the-park…” but NOT “see” things like “danger areas” or “ecclesiastical boundaries” or “where in the oceans a country’s exclusive economic zone begins and ends…” (and I could continue with such examples forever).  This concept of “layering” is related to why we admonish not to “tag for the renderer,” by which I mean that when people map, they like to see WHAT they mapped.  (And so, Carto, OSM’s “front door” renderer, but by no means “the definitive expression of our map (data)” has evolved to render ‘more common, frequent POIs and everyday items” to the exclusion of the more esoteric).  And that’s good, because it means that there is a deliberate filtering going on — a “quality good, noise bad” deliberateness — that helps define what our wider community considers important.  Being able to discern “we believe these data are good and important” does NOT mean that we consider OTHER data as BAD.  It simply means that we don’t render them in a particular renderer.  And so, all kinds of other renderers, routers, use-cases, apps, filters, ways of slicing and dicing our data into whatever we want to see it baked as when it “comes out of the oven” have evolved over the years.  This is correct.

Going even further, I get it that we don’t want the very fabric of our map to be a “garbage dump.”  But what I think is going on, at least partially, is that data being in our map means that we MUST pay attention to them, or be upset with them that “they are getting in my way of what I want to see in any given context.”  Let’s remember that OSM is a shared fabric, and “censorship of data at its very entry” is a method by which we wouldn’t be “sharing well."

> I think that this way of thinking ignores an important aspect of OSM, and that is the shared responsibility for the whole map.

I think this way of thinking ignores that the map is shared in the first place, and if you don’t wish to pay attention to the trees and grass in the park because you’d rather pay attention to the bicycle parking and playground, because that’s what’s important to YOU, how might you determine that someone isn’t mapping well as they add trees and grass to the park?

> It is true that a religious boundary is probably not relevant for tree mapping, however I do have to take other things into account - for example, I could not just switch off the building layer when mapping trees because I might then add a tree inside a building which, while not strictly impossible, is certainly something that warrants a closer look. Or vice versa, a building mapper should not switch off the tree layer for the same reason. Or the roads layer, or the waterway layer, etc.

We might get to a point where “technical layer mapping,” by which I imagine that it is very easy to (technically, in your app, or your browser, or however you see and use OSM data) allows you to turn on and turn off features and objects in our map (perhaps based on specific tags, or themes of groups of them).  We have what I’ll call an early (very early as I imagine future possibilities of using our data in amazing ways) versions of this.  But today, I think PART of why we have this attitude (on one side of the spectrum) of “don’t ‘pollute’ our map with such esoteric data — they feel like NOISE to me!” Is because we don’t have sophisticated ways of “turning this off but that on” in “layers” as much as we might wish.  But in the future, we will, of that I am certain (because I see it happening for the long-term along those lines, and I have for many years).  In the meantime, I think we can get a lot of mileage out of asking ourselves (especially as one might bend towards “don’t add ‘garbage' to the fabric”) to be able to “imagine away” what we might wish to “filter out” of any given data slice of what chunk of the map we are examining at any given time.  I realize that asks to balance longer-term future goals of the map with realities of how it is used (and preferred to use, really) today, and that some might find that challenging or difficult, but it is a longer-term strategy for “dealing with this.”  Yes, part might also be good dialog on the topic of “we REALLY don’t want to be putting THESE sorts of data in our map…” and those ARE important conversations to be happening.  But let’s be careful that what we do with that is a kind of censorship, what in a USA free-speech legal realm is known as “prior restraint” (the attitude of “you can’t film that, or report on that in your newspaper, or say that out loud…” and that isn’t right).

> Already, within the boundaries "layer" there are tons of situations where it at least superficially appears as if one mapper has in their mind "switched off" some layers. For a random example, see: https://www.openstreetmap.org/relation/11687912#map=19/49.00522/-122.76048 - is it really correct that the protected area in the South overlaps with the Indian Reserve in the North? Or is this a case of "layer thinking" where the mapper of one "layer" did not care how their data interacts with other objects?

This is an example of “layer thinking” in a specific case, where the questions are good, and the answers are not readily apparent (they could be, but with examination, wide and deep dialog, and the emergence of consensus).  But it shouldn’t happen in a backdrop of “prior restraint” guiding whether or not such boundaries do or do not belong in the map in the first place:  boundaries are one of those ticklish, odd/strange examples of something we all agree are not readily “on the ground verifiable” (most boundaries), yet we universally agree boundaries belong in OSM (they are here by the million).

> How likely is it that this parish boundary https://www.openstreetmap.org/relation/6854296#map=18/46.84503/-71.22405 is just *slightly* at odds with the civil administrative boundary that runs along the same course https://www.openstreetmap.org/relation/8382360#map=18/46.84503/-71.22405 - or, again, just "layer thinking" a.k.a "I don't care what other stuff is in the map, I want to add MY stuff"?

These questions happen in OSM, they can be solved.  This is one example, there are thousands if not millions of these.  It takes time, dialog, agreement, consensus, wider understanding, but they do get solved (over time).  I don’t think we can take (easy) shortcuts here, and I really don’t want to see the censorship of “prior restraint” take the place of good discussion to resolve such “my stuff with the harmony of what’s more widely OUR stuff” dialogs.

> (I have several thousand such "boundaries slightly overlapping but not quite" examples all over the planet. A few of them will probably be "unlikely but correct", but most will just be sloppy mapping or sloppy importing.)

OSM has always had, still has and likely will have, sloppy mapping.  Sure, let’s do what we can to tighten this up both in the present and future.  But let us NOT include as a strategy to do so “if we’ve never seen these sorts of data before, you may NOT bring them into our map.”  That’s the quickest death-knell that could possibly happen to our map, snuffing out novelty, freshness and creativity as certain as swallowing a suicide pill.

> In my opinion, "layer thinking" means giving up on the shared stewardship of the map, and allows everyone to dump their garbage into OSM with an excuse of "you can switch that layer off if you do not like it". I think it would be detrimental to OSM.

There are balances to be struck here.  Obviously, I’m on one side of the spectrum that is “open” (hm, that’s our first name!) and some are more on the side of “OSM’s data fabric is not a garbage dump.”  If one's ability to “filter layers” were to evolve (rapidly?) to help us (quickly?) show us a new tag we might coin or add, that would be technical efforts very well spent, in my opinion.  In the eruv example, if I (with some additional magic and software support better built into OSM’s smarts and offerings) could suddenly show me (render) an eruv that a weeknight workshop was working to map in an urban area, quickly whipping up a partial- then completed version of our construction, that could also serve as a “final product” as “the map” (of the eruv we worked on).  Even better, it would allow the more mainstream renderers to “ignore” the display of an eruv (as they always have, until the proper tags were coined) and all that would be required is to “mentally ignore” the data that make up the eruv for other mappers in the same area.  It is this latter necessity that I think gives rise to the characterization of “I don’t want to see eruv data in my neighborhood, that feels like garbage data” (and I don’t wish to put words into anybody’s mouth with this specific example) and that is a concern, obviously.  So I don’t know how to “fully solve” that as I imagine in this thought exercise.  But surely, being “more inclusive” towards data entering OSM should guide us before being “more exclusive” should.  Right?

To be clear, there aren’t particulars of “wholly right” or “wholly wrong" here, we’re having a discussion.  I have no hostility towards any particular view.

Apologies for the length, though this is a deep(er) discussion.  Thank you for reading.