[A Rant] Whats the Deal With Crappy Editors and Massively Bad Imports?

I’m kind of pissed off about the import system. It seems like 99% of the imports I have came across are completely wrong. The areas imported often either have incorrect tags and go over large areas that are completely different than the intent of the import in the first place, like the large scrub land area that goes industrial zones and lakes in Calistoga California. There’s also many places in California that I have found where not only was the import is completely wrong, but there were also already multiple other wrong imports of the same exact thing underneath that import. For instance, some of the farm land areas in California have been wrongly important three different times. There’s also places in California where the same national forest has been imported multiple times. It seems like the people California doing imports are either not paying attention or are just importing things in mass to make the map look good when its zoomed out. Most of the large pink farm land north of Sacramento might look good, but they are completely wrong. Also, most towns north of Sacramento is just one large residential area, which is clearly wrong and goes against the wiki rules for the tag. Many of the national forests in Northern California are labeled as parks, which the wiki for the park tag clearly says not to do. Its all rather irritating. Aside from being completely time consuming to fix, a lot of the stuff can’t be fixed in the first place, or it would be really hard to, because the areas are so large and also made up of hundreds of relations that almost impossible to keep track of and deal with. So, I’m wondering what can be done about it. I feel like its pointless to try and deal with the bad areas if someone is just going to re-import things I delete or change tags back to suite them. I’d like to improve the map though, just not waste my life fixing other peoples errors. I noticed on the wiki page about imports that people are suppose to discuss imports now before they do them. Is that really adequate enough to deter bad imports? Why is importing things even a large part of OSM at this point? since there is satellite images and so many things can go wrong, it seems like imports would be an obsolete way to add data to map and would be depreciated like old tags are, especially importing huge areas. Is there anything I can do about this, besides feeling discouraged by the whole thing and possibly giving up mapping?

As a side note, I happen to live in a fairly small town and I am mostly the only mapper here. Although it was an up hill battle due to pretty crappy mapping by people from other areas, I’ve improved things that were already here a lot and added a large amount of new things. The process was definitely more tedious then it needed to be because of mappers in other states and countries not putting the effort into mapping things correctly, because they are not invested in the area, seem to care more about being first place on some top list of “humanitarian” editors, and therefore care more about status than making a quality map. This seems like a pervasive issue. I’ve spent plenty of time reading the mailing lists, diaries (or lately more like add copy repositories), and this forum. I know OSM has problems with bringing in mappers from smaller areas and I know there is a lot of places that are not mapped yet. I don’t think the answer is to create teams of people from other areas that can come in and do the editing on behalf of the local community for points and status. If anything, it just makes things worse. All things considered, I can fully understand why people in small communities would only edit a few things and then never come back, or not even create an account in the first place. Who would want to spend 99% of their time getting over the steep learning curve by fixing other peoples mistakes before they could add things themselves? Especially when it comes to people in small towns, who often have a low tolerance for technology, city dwelling “geeks.”, or people from other countries for that matter. Not to mention, especially when it comes to the areas with the problems I have discussed early, their whole town is already covered by one huge all encompassing scrub land and a giant residential area. So there isn’t anything they could do in their area anyway, except map a house that would be covered in bushes. Well I commend the OSM community’s adoption of more neutral tone and push toward more inclusiveness, I think the issue I have noted here should also be considered as things that potentially push people in smaller areas away. I know if I did not have past experience with computer programming and a love of maps, I probably would have given up a long time. At least something can hopefully be done about the import issues. I’d also like to see the thing about bad editors with big egos from other areas then they are editing be dealt with also. Although I’m not sure what fix there could be for that problem. I think both problems, the importing and the outside editors, partly come from a lack of patience. I am competent the map will be filled in everywhere it can be eventually. Turning the whole thing into a game or making amounts of edits a status symbol in order to do it just seems wrong and counter productive.

Thanks for reading.

There is an interesting discussion on whether OSM should actually be encouraging more imports at https://lists.openstreetmap.org/pipermail/talk-gb/2017-December/020992.html and earlier related threads about Shell garages. After some early bad experiences, e.g. Tiger in the USA, and NAPTAN in the UK, there has been a clamp down on importing and mechanical edits, to the extent that many people are being put off contributing, and others ignore the rules, relying on not getting caught.

It does, seem to be the case that technological hobbies are the preserve of rich countries, and possibly rich areas within countries, which is probably why so much third world mapping is done by first world people, even though GPS capable phones are probably quite commonly available locally.

OSM has to work within the psychology of the people prepared to map, so if people are prepared to put in quantity, rather than quality, and there is a lack of local people, it is going to be prepared to accept those contributions. Unfortunately, no-one really like maintenance work, so initial errors don’t get corrected and subsequent changes aren’t captured. I have mixed feelings on that, having taken on a few local shopping streets which were mapped by a one time visitor with a camera, and knowing that many of their mappings are not maintained at all (I must do a review on those soon, as I know their accuracy is decaying.

Also, you would make a clearer point if you outlined your main points at the start then used shorter paragraphs.

@Adamant1 hadw has already touched on most of the salient points. So just a few additional remarks: California has suffered from a large number of bad landuse imports in the past, that the main actor, while acknowledging that they are less than perfect, continues to take the stance that they are somebody else’s problem.

The good news is that I believe the US community has realized that there is no short cut to data quality and that that now imports are viewed with a fair bit of scepticism. We naturally still have a lot of complainers including those outside of OSM that love to pontificate on why we are stupid etc etc for not making imports “easy”, however the rather low formal hurdles that are in place are really just a miniscule bit of work, compared to actually getting an import right.

All of the above naturally doesn’t really help when you are lost in a import wasteland, the only thing to do is pick away at the problem, and maybe enlist the help of others.

Simon

PS: a couple of paragraph breaks would really help your posting

Thank you AdamAnt1, SimonPoole and hadw. While I won’t address “crappy editors” (except to say that both iD and JOSM have gotten much better over the 8+ years I’ve been in the project), yes, California suffers from many bad, even terrible, imports. Yup, I’ll name names: nmixter, you know who you are.

There is more than one bad actor, but this single one (he is a he) is responsible for MANY low-quality California imports — I have characterized them as “spilling buckets of paint for the sake of doing so.” Over time, he and I have become friends and hiking buddies (he lives two counties away from me), and because my approach to OSM is “quality above quantity over all else,” whereas his is virtually the opposite (whether he admits it or not) I did “manage” an import with him, Monterey County Farmland, a state dataset. My approach in doing so was to show him that data which must be “technically massaged” to comply with good OSM conventions absolutely DESERVE this treatment and “here is how to do it.” Alas, while the import was successful (it took months of careful, step-by-step work, this was intentional), the lesson to this OSM import “offender” didn’t really “take.”

What do we do? Clean up such noise and bad data as best we can. I understand that getting such a rant off your chest feels good, and in a forum like this can allow more light than heat to emanate. But like TIGER, like “Pokémon Pollution,” like other such wrong/bad data in our map, sure, avoiding bad data coming in is best, cleaning up after the fact is second best, so that is what good-meaning OSM volunteers must do.

OSM isn’t perfect. It is noisy, it is messy, it is human, and nobody is perfect. As we embrace it, we embrace its warts, weeds, old data and defects, too. Can we, should we be on the lookout for bad data imports? Of course. Can we, should we clean them up when they happen despite our better vetting to prevent them? Of course. Complaining about it might be a first step (if we must), but the roll-up-our-sleeves effort to make a better map must include not only the blissful editing of adding our most recent hike, the café we just discovered and the park boundary we just learned, but cleanup and updating of bad/wrong data: these are simply part of the territory of mapping.

Imports DO have a place in OSM. For example, very large parks (e.g. national parks, state parks) and specified boundaries of national forests are simply impractical to map via “on the ground” techniques, and so importing such boundaries via a proper source, vetted as to accuracy and sane as to its precision so it meets OSM’s conventions and guidelines are valid data which I consider welcome in our map. If the tags are wrong, fix them. If the boundaries have changed or moved since their import, update them with better, newer data. But please don’t complain that such data don’t belong in our map, as they do.

I heartily encourage any and all organized efforts to clean up bad imports and bad data (even if they simply need updating), whether via wiki, meet ups, Mapping Parties or any combination. This only strengthens our community.

SteveA
California