Changeset: 52813481
CA, around Scotts Valley, remove a bad "farmland" landuse which is actually almost entirely mixed forest and residential
Closed by bdiscoe
Tags
created_by | JOSM/1.5 (12921 en) |
---|---|
source | Bing |
Discussion
-
Comment from stevea
bdiscoe: Please redact this farmland removal polygon, or at least redact your changeset and update the polygon so it is landuse=residential.
If you read https://wiki.openstreetmap.org/wiki/Santa_Cruz_County,_California#Landuse (paragraph six) you'll see that the County does zone every single hectare in this zone to be "Residential-Agricultural" and exhorts OSM contributors to "improve these!"
I don't have a problem with you changing this polygon to landuse=residential if you believe that to be more correct: they are a mix of residential properties which are allowed by county zoning to also allow commercial agricultural use.
But I do have a problem with you simply removing it without replacing it with anything better.
SteveA
Santa Cruz -
Comment from bdiscoe
Hi Steve, since you expressed concern, I took another look at this area in detail. It is a total mess. The so-labeled "farmland" was not just wood and residential, but also some meadow (redundant overlap), and adjacent to existing natural=wood which is, in turn, actually residential in parts, and has overlapping meadow! In short, many of the larger imported polygons are inaccurate to the point of being completely useless and inaccurate. What's worse, their presence discourages anyone from actually putting the correct landuse in there, because a new correct landuse would get tangled with the old incorrect landuse. The one farmland I removed is only the tip of the iceberg of what needs to be wiped and re-created here. Do you have a better suggestion?
-
Comment from bdiscoe
As for "removing it without replacing it with anything better", if you have a large area marked "D" which is actually A, B, C, then removing the wrong "D" IS actually better, because it is actually less wrong.
-
Comment from stevea
I don't think there is anything terribly wrong with meadow overlapping with other landuse, and nobody has said so in Santa Cruz County, where we have been doing this for at least 8 years. Look at Wilder Ranch State Park, what many have called "visually pleasing" (meadow overlapping with landuse=forest, landuse=wood and leisure=park). The history is complicated, the rendering is complex, it has evolved to be this way over quite a long period of time and with a great deal of discussion and consensus.
Have you READ our County page, as I have recommended? There is quite a bit of history here, by locals, using local data, over many years — for most of the history of OSM as a matter of fact. I don't edit San Ramon (though I know Danville, Pleasanton and Livermore fairly well and have mapped there), but to simply redact official landuse data and berate local convention doesn't sit well with many here.
Regarding removing a wrong "D" well, I'm OK with it if you can justify doing so (and you haven't, except to say "it is a total mess.") So that doesn't fly right from the start. But you didn't replace "wrong D" with A, B and C which are correct. If you do so, I'm OK with that. But you haven't. Since you use the word "re-created," I await your re-creations. Do you intend to recreate A, B and C or will you just redact without any improvement? If the latter, I will restore the polygons, perhaps this time as landuse=residential instead of landuse=farmland, as our local wiki suggests. (It predicts that they may oscillate between these, they have and they do, as contributors indeed "improve" them).
Why you would say that larger landuse polygons are "inaccurate to the point of being completely useless and inaccurate?" Our wiki states "consensus (has) emerged of capturing zoning with landuse=* is a good first step to avoid large blank areas, but when actual on-the-ground data are also known, they are preferred to simple zoning (landuse=*)." DO YOU have any actual on-the-ground data to contribute to IMPROVE what you have redacted? If so, do so. Otherwise, please leave our perfectly valid "good first steps" in place. The notion that these discourage better data is nonsense: rather, you (and/or other OSM contributors) appear to be lacking in better data with which to do so.
SteveA
-
Comment from bdiscoe
I have now spend some time in the Happy Valley area to expand the natural=wood relation down from the hills to fill in where there is actually wood, which is most of the area left blank by the "farmland" removal. I also validated the relations to solve overlaps and degeneracies.
-
Comment from stevea
Well, thanks for that; I'm watching. It's possible we posted Comments so temporally close together that we crossed each other, but I do await answers to my questions. Thank you in advance. SteveA
-
Comment from stevea
However, if you "expand the wood" (e.g. as it appears you have from the east side of Happy Valley Road to its west side), you truly break many landuse semantics as published by our County GIS: landuse=forest really is timberland, natural=wood originated from "special_use" polygons which have a specific purpose but which are largely wooded. Do you really believe that you can use Bing (and a guess and a prayer) to better define landuse than does our County GIS department?
-
Comment from bdiscoe
"you can use Bing (and a guess and a prayer) to better define landuse than does our County GIS department?"
If the County is marking forest as non-forest, and non-forest as forest - which they clearly are here - then absolutely. Detail based on aerial is based on detail that is just wrong.
-
Comment from bdiscoe
(...is better than...)
-
Comment from bdiscoe
Also, somebody foolishly imported fields like "Shape_area" which will be wrong the moment that somebody comes along to fix the polygons. What on earth is the point of that? Surely nobody expects general OSM users to update the useless "Shape_area" when the move the nodes into a more correct version? I strongly recommending deleting these spurious "Shape_*" tags. Also, "attribution" and "source" tags belong on the changeset, NOT the way; they next person to come along is not going to change every since "attribution" tag, so they have no purpose except to be soon wrong.
-
Comment from bdiscoe
Yes, I have read the "County Page". As for "nobody has said so in Santa Cruz County, where we have been doing this for at least 8 years", if nobody has pointed out that this data is mostly wrong, then it's only because nobody has looked, which is highly believable since there appears to be no actual data here, ONLY IMPORTS. Look at the Tiger ways! E.g. "Happy Valley Road" where the nodes are up to 100m off. 100m! Aligning the Tiger to reality is among the first steps to do in any part of the USA. It's the first thing I did in San Ramon. Sadly, nobody has done it here, which would explain that nobody has also noticed that the landuse is largely fictional.
-
Comment from bdiscoe
My recommendation:
1. Either remove all the bad landuse, or use a JOSM filter to hide it for step 2.
2. Align the roads and fix the road topology!
3. Only then, you can gradually start to add landuse, carefully, bit by bit, checking and correctly each batch before upload. -
Comment from stevea
bdiscoe: Your hostile and hyperbolic comments that "there appear to be no actual data here, ONLY IMPORTS" and "largely fictional" are unwelcome, untrue and show you to be a histrionic exaggerator. I have been mapping this county (lovingly, carefully, with my GPS, notebook and tens of thousands of edits) for over eight years. PLENTY of people have looked! Santa Cruz actually won a BestOfOSM.org award for having "nearly perfect landuse!" Earlier this year, I was awarded OSM's Mapper of the Month award. I have spoken a number of times at SOTM-US national conferences. The university (my alma mater, where I work with many) uses OSM as its basemap. I constantly hike and bike yet more and more remote areas of this county and then generously contribute the highly accurate data to OSM.
What you propose is fantasy wishful thinking: you say "fix the road topology" like a magical super-edit is going to come along and suck the knowledge/data out of the ether and plunk it into our map. Nonsense! We have county landuse and we have on-the-ground data (over many years, refined with many versions) and both together are pretty good (we win awards, speak at our national conferences and are named mapper of the month) and they CONTINUE to improve as I and others around here hike and bike yet more.
You haven't answered my questions: what is the source of your better data? Do you live here? Is this your backyard? Have you ever even set foot in Santa Cruz County? I have lived here most of my life, have hiked and biked most of this county and mapped much of it, with the help of many others.
I'd like to invite you to stay out of editing in this county: your hostility and attitude are unwelcome, your data is NO better that what we have here (you continue to fail that you can't prove that) and I'd like this comment thread to come to an end with your declaration that you might not like our data, but you haven't any BETTER data to contribute, so you won't edit here.
Kindly, SteveA
-
Comment from stevea
Regarding "fix the road topology," you will see (if you look, try http://product.itoworld.com/map/162?lon=-121.91947&lat=37.02996&zoom=12&fullscreen=true) that NOBODY has fixed more TIGER misalignments in this county besides me. By that Ito map, it is one of the most "blue" (corrected) counties in the entire state of California.
Our wiki even explicitly states that road are up to 100 meters off! Why are you so shocked that they are! This is not my fault, it was a messy TIGER import and we all know that. Yet, in eight years there are still some roads which are so remote and/or closed with gates and/or on private property and/or invisible from Bing or other imagery that they are nearly impossible to correct using those methods. Does that stop me from intending to do so? No. Does that mean that they should be wholesale deleted? Well, no again, yet I agree with you that these data should be corrected. But correct them with WHAT? You do not appear to have better data which superseded our already "OK" or "fair" data. So, keep your crayons in their box.
-
Comment from stevea
The answer to who "foolishly" imported the SHAPE... attributes is nmixter. He has been hugely admonished for a very sloppy import here, and I have literally spent YEARS cleaning it up. EVERY SINGLE ONE of the >3000 polygons he imported was visually and personally redacted in JOSM by me in the "version 3 update" in 2014. If you are signing up to improving ALL 3000 polygons in our county, I'd like to see your plan to do so. But to randomly "nip at edges" that this one or that one is wrong, without fully understanding the history of the data that are here, you do yourself, OSM and our data in this county a huge disservice.
-
Comment from stevea
I wait no longer to do what I have known to be the right thing to do since this dispute began: I'm redacting bdiscoe's polygon removal edits.
Sadly, and I have never done this with any other OSM member with whom I have "friended," I also remove him as my friend in this project. His arrogance and hostility demonstrate themselves in his comments above, his User page boasts how he is more concerned with what are his "ratings" in OSM's "rank spreadsheet." These are not my values in this project, nor are they tenets of OSM.
Perfect data cannot be the enemy of the good, and good data cannot be the enemy of none whatsoever.
Adios, amigo!
-
Comment from stevea
In the spirit of "pour calming oil over troubled waters," I offer that this dispute largely resulted from tangles among the complex issues of landuse and landcover. They are misunderstood, confusing and fraught with ambiguities. There are a plurality of tagging strategies and histories, not everybody is familiar with how fuzzy are the edges.
Landuse might be residential or farm or even both, as defined by how the property may be and is used (in a zone, in a legal sense). Landcover (like meadow, even though it is specified with a landuse tag) is different — or is it? Forest and wood tags make this even more complex. And rendering in mapnik/Standard and Carto are different still, with issues that help explain certain evolutions, allowing us better understanding of the complexities and histories.
There exist tagging strategies which are correct according to many, have evolved, yet still contain errors in the opinions of some and which others find confusing without some explanation or historical context: all at the same time. Our map is quite plural.
Landuse and landcover untangle slowly, it appears. Misunderstanding that what is seen in Bing as landcover doesn't automatically supersede existing correct landuse. It isn't one or the other and sometimes is both.
We do our best here, OSM evolves.
-
Comment from bdiscoe
So, just to wrap this changeset discussion; it spun out of control to attempt discussion of a lot of things besides its subject, the changes herein. In cases where the local preference is for some unusual interpretation of OSM's tags, it's advisable to just leave it as-is. There is enough to fix in the rest of the planet.
-
Comment from stevea
In this controlled discussion (on my part), the changes herein WERE indeed the subject: it appears not listening to them nor parsing their consequences is one of the blithely scattered activities of bdiscoe as he attempts to keep his scores high on the leaderboard while he preys on other under-mapped parts of Earth.
There is no "local preference for some unusual interpretation of OSM's tags" in Santa Cruz County: indeed, as noted by me and ignored by bdiscoe, landuse tagging is very well explained in our wiki as to its history (and thousands of corrections over many years), has gone through three revisions of improvements, and has won awards.
Instead, it is bdiscoe's apparent lack of understanding (or unwillingness to extend it) that landuse, natural and leisure tags are complex and can (and do) cause confusion among many OSM editors, wherein some think that because of what they see in Bing (landcover), those tags should supersede.
Again, I'll make my point (which bdiscoe never addressed): if bdiscoe has BETTER data to add here, he is welcome to do so. THAT is a tenet of OSM. However, he does NOT have such data and he wants to keep upping his stats, so that is the real reason he does not enter better map data here. Don't be fooled otherwise by what he calls a local unusual interpretation. We're just fine here, and always improving our tagging. Anyone remains welcome to do so, although a requirement that you bring to the table BETTER data does remain a requirement. Also required is basic knowledge between landuse and landcover, the root of bdiscoe's misunderstanding.
Ways (7)
Relations (1)
Welcome to OpenStreetMap!
OpenStreetMap is a map of the world, created by people like you and free to use under an open license.
Hosting is supported by Fastly, OSMF corporate members, and other partners.
https://openstreetmap.org/copyright | https://openstreetmap.org |
Copyright OpenStreetMap and contributors, under an open license |