[Tagging] Fuzzy areas again: should we have them or not?

[stevea]

On Dec 21, 2020, at 12:36 PM, Anders Torger <anders at torger.se> wrote:
> Actually it seems to me that thinking about several tags at a time seems to be a fairly new concept ;-)
> 
> Joke aside, describing a complex reality may require something more than the simplest possible data model.

I must not "get" the joke.  "Thinking about several tags at a time" is something I and others do in this project somewhat frequently.  See our Park boundary proposal [1], for example, and understand how it has several other tagging proposals associated with it (by Brian, by Kevin, by me...) that are helping to advance a fairly comprehensive "park reformation" effort in OSM.  This hasn't been, nor is it easy, by any stretch (I've been working on some version of this for over ten years).  It requires literally years of thinking, planning, design, collaboration, emailing / dialog, writing / documentation and consensus.  Yet we do it, and it is getting good, positive feedback and traction from those in the community who are vocal about it.  This is moving into the realm of "a win for the team," but we're not fully there yet.

Note:  being vocal in OSM, especially in the context of what often, to many, feels like "a harsh (pseudo?-)meritocracy-based lions den of strong language and attitude" is indeed a powerful way to initiate and project serious influence in the project.  Yes, we can and should continue to find ways to be more inclusive to more voices with varied perspectives here, without intimidation or outright hostility.  I strive mightily to be open, welcoming, listening and inclusive (perhaps I miss the mark sometimes, we are all human), and I expect others to be this way, too.  I have, do and will call people out (gently, kindly, even privately if I can) if they don't, as I think we should.  There are good and better ways to do this.  It is bold of anybody to do so, I wield this carefully, as it sometimes it is necessary to moderate and buffer with friendly, better attitude more inline with our community standards.

> The thing got me baffled from the start is that there's so many things regarding naming that can't be done in with OSM and any of its most popular renderers. That have existed in regular maps since forever. To me it's like starting to use a new word processor just to discover that the character 'q' is in available, because the designers didn't put it in as it's not really used as much as the other characters :-). I just "WHAT? You must be kidding!". The omitting of obvious features (what I thought was obvious at least) and the lack of interest to find solutions for them just felt unreal.

OSM evolved.  It wasn't born, and it hasn't become, what you imagine it to be.  OK, please accept this, and / or getting over your bafflement.  It is what it is:  a dataset that constantly strives to improve itself.  You are welcome to improve it using the methods by which we improve it.  OSM wasn't created to "be just like the maps you know" and THAT is why you are surprised.  So, please stop being surprised about that, now that you explicitly know.

> But now after being on this list for a while I'm doubting myself, maybe this isn't as obvious after all

Whether obvious or not, I hope at least that one point is made clear to you now.

> I know mostly Swedish and to some extent Norwegian maps, maybe all these features are non-existent in all other countries, but I doubt it. What is somewhat special is that we have quite a lot of rural areas, but that exists in other countries too. I would guess that there are many native American names still alive and well in the US nature, just like we have lots of Sami names in nature Sweden/Norway/Finland. I think it's more about that most OSMers are interested in urban areas, street routing and stuff like that, and outdoor maps haven't really been much of a thing other than for simple illustrative purposes.

You speculate, please be careful doing that.  I'd speculate your speculation is at least partially correct, but it is not the whole-story explanation of what is deeply complicated (and you continue to attempt to oversimplify).  I started mapping here in 2009, primarily "outdoors" (do we map indoors? — another topic for another thread on another day) as a hiker / biker and "human being out-and-about my shopping districts, parks, beaches, buildings and theaters..." who wanted to improve a sketchy map that needed improvement.  That's OSM in a nutshell for many, but of course there is much more.  People who discuss on this list look to deeper (structural, future-building...) aspects of the project, both asking and imagining / sketching "something better, perhaps."  Please don't think that you can guess that "outdoor maps haven't really been much of a thing..." as that's a guess that (speaking personally) is NOT "a thing."  I assert many others agree.  There are many, many examples of "specific-to-outdoors" renderers based on OSM data, right now.

> I don't consider it a great achievement in design to simply just strip away all features of maps that require a (slightly) more complex data model. What we in OSM seem to have done is simply to adapt the end uses after our data model, not adapt data model after what typical end uses require. It's like if we considers these names in nature to be poorly designed (no defined borders! booo!) and should therefore be disregarded. But this is our culture. Those names exist and any serious map maker takes these names seriously and find a way to represent them. Actually I even find it disrespectful to the people of the local land current and past to ignore the names. Locals will still know them, but visitors will only know them from maps, and if they aren't there, they won't exist to them.

I don't think OSM consciously, explicitly "strips away" features (or tagging).  On the contrary, I think OSM does nothing but become more and more rich with syntax to describe our complex world as map data.  The "tension" between the goal of the map ("let's enter the best map data of earth into an open project") and its end uses will always be with us:  it is impossible to perfectly predict (with our data entry) what end-users are going to want to do in the future (and that IS what you area asking of OSM right now).  So, "next best" is both knowing that impossibility, while "doing our best as we do things today" (with room for improvement, of course).

"Disrespect to locals" seems over-the-top.  While I strive mightily to help you achieve your goals of natural naming (as I and others here understand them — and we continue to offer you our earnest assistance), there has never been (I speak for myself) any wanton disregard or active oppression that would insult locals with deliberate omission or "let's frustrate efforts at Swedish natural naming."  That's simply false and "flips the script" in a most unfair way.  If you want to improve your local area in OSM, do so.  If you are having problems "how," ask good questions without complaining at how things are, with an eye towards improving things so you (and others!) can benefit from the fruits of your efforts.  That is how a collaborative, open data project works.  Saying "we" insult "you" because you are frustrated with our data, or how we model it, is not going to make you any friends here.

> But I hear what you are saying, and many other say. It boils down to that OSM is not a map, it's a database with geo information, and it will only cover those features that fit the data model that is already set (in stone perhaps?). If a map feature needs something more, well, then we don't do those maps, or someone else could add some extra database and make their own renderer that can. Well, I guess that's fair enough.

Better, Anders.  Yes.  But saying "only" isn't so.  That makes it sound like we never grow, patently false.  So does "set in stone."  But you have to understand that a going-on-two-decades-old data model, having grown through sane, thoughtful evolution by thousands and millions of people, isn't going to "turn on a dime" (quickly and radically alter course in a new, different direction) because you are confused when you compare OSM to your national maps.  We can see that, can you?

> I'm today leaning towards stopping to contribute to OSM as the type of mapping I'm most interested in is what OSM on the whole is clearly least interested in, and not really interested in improving in either. But I'll think about it the coming weeks. As Tomas pointed out you can work with OSM without ever caring what people in this list think or do. It basically means that you need to implement a renderer on your own though and make own data on the side and that's a huge task, and I'm not sure I'm *that* interested in mapping.

Please do "think about" (things as you describe them and as I and others have responded).  We're here when you have cogent, thoughtful questions that don't include insulting how we do things today.  And we'll do our best to help you.  You can do a lot of mapping withOUT making your own renderer (I haven't, for example and I find contributing to this project enjoyable and valuable as I "give back").  Perhaps you can find your way to contribute to OSM and find value in return — I hope you do.

SteveA

[1] https://wiki.osm.org/wiki/Proposed_features/Park_boundary