[Tagging] Basic cartography features missing, why?

stevea steveaOSM at softworkers.com
Mon Nov 9 20:03:09 UTC 2020

On Nov 9, 2020, at 12:59 AM, Anders Torger <anders at torger.se> wrote:
> Hello Steve, I admire your passion.

Thank you, that is a word both used about me in this project by others and one I use myself.  Most apt; some of us in OSM are almost "rabidly" passionate!

> This is my perspective: there are many projects to contribute to, I have multiple interests, I have a limited amount of hours to contribute. I have worked in many high risk ventures, and seen at least 10 years of my work life's production has just disappeared in projects that did not make it. I have experienced work-related burnout once in my life, 16 years ago, and I still have not recovered 100% probably never will, so I need to be careful before I bit over something too large.

Burnout happens.  Thanks for sharing that, it offers perspective and potential danger of what can happen if we are not careful to manage our time and risk, even in OSM.  I myself have suffered burnout in OSM, where I have to back away from it for a bit and do other things.  But I do always seem to come back to OSM, it is both rewarding and relaxing, an awesome combination.

> So I make a "due diligence" of projects before I commit. For OSM I actually started before looking into it, fixing roads so the route planner I used for sharing my bike rides would actually work. But this year I've been thinking about stepping up the game, considering mapping, not programming (that's a whole different level of brain activity).

For "sharing bike rides," some of my compatriots / associates in the world of mapping and biking find that RideWithGPS (.com) provides the needed services; check it out.  (I'm not a member nor do I get anything for saying that, it's simply a recommendation I can make as people I know say that it works and I've used it and tend to agree).  I don't know if it works in Sweden, maybe, maybe not.

Coding is different than mapping, that's for sure.  Both are cerebral tasks, there are different left-brain / right-brain mixes in each of them in different ways (brain hemispheres).

> Then I first started with testing if OSM feature set was currently capable of generating maps to a similar level I'm used to from the real maps I've been using all my life. It was not in some important key aspects, and I was a bit surprised of the type of features that was lacking as I consider them central for good cartography. So I started investigate how these could be added, first simply by reporting them as bugs, with an initial suggestion how they might be fixed. But long story short, I've come to realize that it's a multi-year process, and the interest of having these "basic" cartography features is weaker than I expected. I've also investigated import status in Sweden, why the situation is like it is despite five years of good data being available which could significantly improve the baseline. The result of that investigation is that we don't have much if any organization locally, and it's technically hard to make imports, those that have tried have got less than optimal and/or only partial results. I've also looked at community statistics. There are about 50 active mappers in Sweden. Despite me being rather casual mapper, I was recently placed top 15. And we have the competition from easy access of high quality maps. This is a tricky situation for any project.

Anders, I (for the third time) have gone back and read your "list of nine" things an earlier message of yours in this thread states are "limitations" and I'd like to ease your frustrations about those.  I'm not quite sure whether we should be doing this to the wider list, so perhaps we take it off-list, as you have specific concerns that I'm not convinced will benefit the entirety of the tagging discussion list.  But briefly, you seem to have both specific concerns about "mapping limitations / cartographic features missing" and what might be called broader OSM-wide concerns about the project as a whole (professionalization, imports...).  I feel I could more easily help you with the specifics of "limitations," (as I don't believe you are really as limited in mapping as you believe you are).  So, let's take that discussion off-list.

For example, you complain that natural=peninsula doesn't render.  So?  That's not a problem with OSM, it is you assuming that a particular renderer is going to display the semiotics you believe it should, when it likely does not (exactly as you believe it should).  That doesn't mean OSM has "basic cartography features missing," it means you are expecting something from a particular renderer it likely doesn't perform or deliver as you expect.  That's all — not that OSM is broken.  And I still don't understand what your complaint is about "tagging and naming areas 'on the ground' does not seem to be developed much at all."  I'm tempted to say "nonsense," as I certainly don't have this problem (nor do millions of other mappers, it seems), so I'm curious for you to describe what your problem is here in more detail.  In fact, most or all of your frustrations seem to be about what "a" or "the" renderer displays to you.  That's not a complaint you can level at OSM and its "basic cartographic features," although I understand your frustration and continue to listen.  I'd like to address what CAN be mapped in "your world" and have the results work for you, THAT is what makes OSM a success in most people's minds, seeing the results of their mapping on a map.  But it doesn't always happen as people expect, especially when they appear to be mimicking a very richly displayed "gold standard" set of semiotics from another style of map.

> It seems like due to the limited maturity locally, I need to go in full time and become an OSM champion, and that is what you tell me as well. Indeed I think I have the competence (or rather have the basic skill set so I can build it), but unfortunately not the capacity, and if I need to spend a lot of time I prefer projects with more programming and less politics. It's natural that large communities have lots of politics, so I tend to prefer smaller projects, or alternatively large ones with clear roadmaps and well-defined sub-tasks to engage in (my Linux kernel work belongs to the latter).

Heck, if s/he gives me half a chance to do so, I'll tell almost anybody who'll listen to me to become an OSM champion — and you have!  I wouldn't call what we're doing "politics," nor do I believe that "politics" are required for you to do the mapping you want to do.  Although it does seem that a particular direction you want to go in is entering map data from your government's "gold standard" maps into OSM.  Be careful doing that, as one of OSM's important tenets is "don't copy from other maps."  Sure, I get it that you want to "richen up" Sweden in OSM, and that's terrific, but not necessarily by directly copying from government maps (in bulk, as a giant, "whole data" import).  Sprinkling in the name of something from them here or there, that's different and likely OK, so go ahead and do so where appropriate.  But that's not politics, it's simply following our ODbL and "not copying other maps wholesale."

> If I wait for the project to move on its own here locally, I consider it to be a substantial risk that nothing will happen, but who knows, hopefully someone has the skill, time, passion and capacity to engage. It also seems like cartography to the level I expect is not a key priority within the majority of the project, and to me personally that takes away a big part of the joy of mapping. I don't just want to make a geodata database, I want to see beautiful useful maps. Of course all want that, but there are many different views on what makes up useful maps. To me the generalization shortcomings are just too large, and have been that for a long time.

I wouldn't lose sleep over what appears to be "a substantial risk" that you (or anybody) not entering data in a particular area is a clear and present danger to the map.  It isn't (substantial), but it is true that "empty deserts" (of OSM, with little or no useful data over a wide area) do present SOME risk to the future of the project.  However, like I said, "the only way through, is through."  There are no shortcuts or magic wands to data entry in OSM, it simply must happen, yet it takes time and effort.  There are better-established strategies, like keeping an eye towards high quality throughout, starting with "basics" like roads and waterways (the "bones") and using these skeletal structures to build other "organs" of what makes up a map, over time.  I disagree with you that "generalization shortcomings are too large and have been for a long time."  There are millions of us mapping who simply don't find that to be true.  And I'm not sure what is blocking you from grasping that, so I'd like to (passionately, yes) help you overcome that.  Again, OSM is about the data in the database, not any particular rendering of a map.  Please do not confuse the two.  This can be a difficult concept for many to grasp and even difficult to communicate to people (which OSM does with reluctance, because it can cause them to "become lost" and lose sight of the "reward" of seeing their mapping efforts "blossom" as they enter data).

> About overall maturity, I don't think 16 years runtime makes a project automatically mature. Another perfectly realistic scenario is that the a project outgrows the initial processes making it stagnant in certain aspects which causes various functionality gaps that wouldn't be expected of a project of this age. It's my assessment that this is exactly the current situation with OSM. No assessment is perfect and 100% complete, but for my own purposes I've got enough of those indications to come to that conclusion. This doesn't mean that it cannot change, but it's not something I can do. The only thing I can do is to report what I see and then the rest of the community can do whatever they like with it.

OSM, being "16 years old" isn't an automatic "well, if it is that old, it must be mature," I only said that because 16 years old, in the case of OSM, IS "old enough" to have developed a great deal (certainly not everything — "the map is never done") of the tagging and infrastructure it needs for people (like you) to map, and even map quite well.  It appears you continue to be frustrated with what you identify as "basic cartography features missing," yet I'm not convinced of that.  I still think there are tags you might use (or coin) that can create the sort of map data you believe should be in the map.  Though, 16 years is also sort of "adolescent" or even "very young adult" stage (for human development), I'd even say that OSM is in a similar state of adolescence or young adulthood, and therefore in an interesting state of (early) maturity.  Maybe OSM's lifespan / lifecycle has a timeframe similar to humans:  I wouldn't be surprised if OSM or something it might morph into is around in fifty or a hundred years from now.  OSM is simply a darn good idea and has become seriously established.

> In all, for me probably the wisest thing to do is to step back and continue my casual contributions as needed for my bike routes, and keep and eye and see what happens with OSM at large and locally in Sweden, but not engage in detailed cartography until it can be represented well.

There you go again assuming that because "it isn't represented well" (in a particular renderer) OSM is at fault for having "features missing."  Not so.  I agree that this can be a concept that many struggle with, but I ask you to please separate these issues in your mind, as it resets your expectations properly.  Entering correct data into OSM as a database is the goal, seeing these data render in a particular renderer is a different task only tangentially related to the data being there.

> Regarding suggesting things, I think I have made plenty of suggestions, both in terms of how features could possibly be implemented, and how priorities could be changed etc, and many of these are just the same as many others have suggested in one way or another. Of course, I cannot expect these to be implemented just like that, and obviously many think that these suggestions are outright poor ideas, perhaps the majority of the community. I don't know. This is just a view of some random guy making a due diligence of the OSM project from a Swedish casual mapper perspective:

In the list that you provide (excellent!), let's break the components into two broad classes:  those that are more "organizational" suggestions (expert groups, how to organize imports, community growth, marketing / "selling," OSMF board "roadmap,"...) and those that are specific to mapping issues (which you characterize as "basic cartographic features missing" which I don't believe is true, but rather a misunderstanding of how to enter data into OSM and perhaps see it render on renderer-A or renderer-B, perhaps not).

As I'm not on the board nor do I have any more "pull" (influence) over how the "higher level" tasks you identify play out in the future, let's subtract those (for now) from this tagging list discussion (as they aren't really appropriate here, though they are concerns of yours and should be addressed, just not here).

For the specifics of "how do I map these features?" concerns you have, again, let's take that off-list.  Those would be appropriate to discuss ON list, it's true, and maybe you publish the RESULTS of our off-list discussion here after we've emailed each other.  But I feel we have spent a great deal of time (and passion!) here on this list and it's getting tedious for other readers.  This thread is LONG!


More information about the Tagging mailing list