User talk:Vid the Kid

From OpenStreetMap Wiki
Jump to navigation Jump to search


Yes, I cringe every time I get to that part of the state in OSM. In some areas just to the east of Dayton, it's not just nodes that are duplicated, but entire ways as well. And in Butler County (Hamilton area), a few of us have been busy correcting every single way in the county. It looks like someone was drunk when adding that data to TIGER! :^) Working on Central Ohio last night was a pleasure by comparison.

Thanks for your help on the route relations.

 – Minh Nguyễn (talk, contribs) 21:53, 14 April 2009 (UTC)

Interesting: the entirety of Greene County is duplicated. Soon after DaveHansen imported the TIGER data nationally, RSatterf went in and imported all the Greene County data, which means every single way in the county is duplicated. [1][2] RSatterf is also responsible for the countless "loose" nodes all over the county. I'll contact the mailing list to see if anyone has a cleanup script. – Minh Nguyễn (talk, contribs) 22:04, 14 April 2009 (UTC)
Okay, Dave's helping to clear out the duplicate data. [3] – Minh Nguyễn (talk, contribs) 02:27, 15 April 2009 (UTC)

Re: SR vs US:OH

Oh cool, Mapnik now renders refs with colons in them, like US:OH. I used to tag that way, but the renderers refused to display anything, so I figured the SR notation was at least more useful for the time being. I'll stop retagging state routes that way, then. Sorry about trampling over your work.

Currently only OpenCycleMap, which is based on Mapnik, recognizes relations. That gives me hope that it's easy for the other renderers to support relations too. But I'm not even sure if OpenCycleMap can handle concurrencies well. After this weekend's downtime, I'm going to start tagging the Ohio-to-Erie Trail, which runs along the Little Miami Scenic Trail (rcn 1) through Southwest Ohio. We'll see if it works.

In terms of transitioning from refs on ways to refs on relations, nearly all the ways are tagged with a state (even with the SR notation), thanks to the multitude of tiger: tags. So once I get the hang of the extended API, I'll start creating relations based on them.

It seems like there's a consensus that county route relations should be tagged ref=US:OH:County. For Ohio, there's actually three styles of county routes: most counties don't sign their routes, Jefferson and Ottawa counties use the MUTCD blue pentagon, and Harrison County uses a white square. Since they really are separate networks, shouldn't we be putting the county name in the network, not just "County"?

I'm pretty excited about the prospect of finally using custom shields. I happen to collect official state and county highway maps, and it seems they're the only ones that ever use the custom shields. If we were Kentuckians, it wouldn't matter so much: they use the generic circle shield except for special roads like the AA.

 – Minh Nguyễn (talk, contribs) 08:17, 18 April 2009 (UTC)

Mind if we move this discussion to Talk:Ohio, where others may notice it more easily? – Minh Nguyễn (talk, contribs) 09:21, 20 April 2009 (UTC)

Go right ahead and repost. I don't have a problem with it, and I couldn't stop you if I did. :-) Vid the Kid 21:59, 30 April 2009 (UTC)
Discussion moved to Talk:Ohio.


Oh cool, OpenCycleMap handles concurrencies just fine: this is the Little Miami Scenic Trail in Loveland, where it's State Bike Route 1, the Ohio-to-Erie Trail (ODOT Bike Route C), and the Underground Railroad Bike Route (proposed U.S. Bike Route 25). At lower zoom levels, OCM switches to badges; it looks pretty random to me, but at least all three routes' badges show up at some point or another. – Minh Nguyễn (talk, contribs) 00:29, 30 April 2009 (UTC)

Yes, I've heard OpenCycleMap makes use of route relations, at least for the bike routes. I would expect that any renderer that pays attention to (road) route relations, and is implemented in some way that's not totally stupid, should do overlapping routes just fine. We just need to get Mapnik to do that. (Forgot to sign) Vid the Kid 21:57, 30 April 2009 (UTC)

County borders

I've done some work with county border relations in Southwest Ohio, as you mentioned, but I've only completed the work for Warren County, so are still double borders. It's a pretty tedious process, and it doesn't help that some users have begun importing city borders without joining common nodes. (So where two cities border each other along a road, there are three points atop each other all the way down the road.) Northern Kentucky's city boundaries have been imported already; not sure when they'll get to Ohio. – Minh Nguyễn (talk, contribs) 17:53, 25 June 2009 (UTC)

Yes, I see now that the border ways (at least, around Warren County) are indeed merged. Unfortunately, Mapnik still draws two overlapping borders, one for each county. (Osmarender doesn't seem to render them at all.) Vid the Kid 21:36, 26 June 2009 (UTC)
For Osmarender, the issue is my use of admin_level (per the wiki documentation), not relations. It only starts displaying admin_level=6 boundaries at zoom level 14, whereas Mapnik displays them at 11. Annoyingly, neither renderer shows the admin_level borders at zoom level 9 like the untagged borders. Also, Osmarender still won't render landuse names until the very last zoom level. Grr. – Minh Nguyễn (talk, contribs) 02:59, 27 June 2009 (UTC)

Re: Route 4 != Dixie Highway

Alright, sounds good to me. – Minh Nguyễn (talk, contribs) 04:17, 1 July 2009 (UTC)

Expressway page

Thanks for writing up User:Vid the Kid/Expressway. That's about the most coherent explanation I've read so far on this topic. High-speed, non–access-controlled expressways are a major concern in Southern Ohio and Northern Kentucky, where the Appalachian Highway and AA Highway are the main expressways. Tagging them as motorways would lead users to think they're as safe as Interstates, when in fact they're much more dangerous. Kentucky has taken lots of safety precautions on the AA Highway, because people treat it as an Interstate and then hit a sudden stoplight, causing frequent accidents. Going one step down will remind users that these routes require a bit more caution.

Like you mentioned, we definitely need to distinguish from "conventional roads". On the Appalachian Highway and plenty of West Virginia's highways, most junctions are at-grade intersections with nothing but a couple signs. A user planning to travel on a cross-street should be informed that the road will have a rather hairy intersection with this expressway. It's not just rural areas, either. There are a few notorious intersections in Hamilton County like SR 126 at the off-ramps for I-71, which is why I took care to use highway=trunk in those areas.

If OSM ends up using the HFCS, I hope we take your considerations into account.

 – Minh Nguyễn (talk, contribs) 08:49, 24 September 2009 (UTC)

Re: US 42 is apparently north/south in Ohio

Okay, I wasn't sure how much to trust the directions given in the straight-line diagrams, since they indicate I-275, a loop highway, as being completely east-west in Ohio, which isn't correct. (At some point in Clermont County, it switches to north-south. I didn't get a chance to verify in person the last time I visited Cincinnati.) – Minh Nguyễn (talk, contribs) 10:06, 24 September 2009 (UTC)


Thanks for taking on the CDP work. It's odd that whoever imported these TIGER boundaries didn't bother to upload township lines, which are also recognized by the Census Bureau. I can fill in some of the township lines in the Cincinnati area, based on maps like this, but we need something larger-scale. Also, the TIGER placenames for townships don't include "Township" the way they should in Ohio. – Minh Nguyễn (talk, contribs) 10:41, 21 October 2009 (UTC)

Maybe the census bureau didn't include township lines in the particular data file used by the cities import. Or maybe the person who scripted the import filtered out townships, believing them to be unnecessary clutter. (True of CDPs more than of townships...) Anyway, I think it's obvious that townships should be admin_level=8. I also think that, because townships have no jurisdiction inside incorporated areas, they essentially don't exist there. That is, they should be mapped as not overlapping incorporated municipalities. Agreed? Vid the Kid 10:52, 21 October 2009 (UTC)
Apparently Ohio has a quirk allowing a township and a city to share jursidiction over a piece of land. I said as much in this Wikipedia article. There are three instances of dual jurisdiction in Medina County, which I added to the article. Loveland and Symmes Township in Hamilton County were considering doing that to the high school, to resolve an annexation dispute. But I agree with using level 8 for townships anyways, since these cases are rather rare. Lordsutch, the city line uploader, comes from Texas, which doesn't have township governments. I've asked him whether he could add the township lines in. – Minh Nguyễn (talk, contribs) 06:13, 22 October 2009 (UTC)
A second import for the townships would probably result in a lot of duplicated ways where municipalities and townships meet. (Not to mention along county lines...) And, depending on the format of the data source, the townships could have holes for every CDP and would probably not include those rare instances of dual-jurisdiction you mentioned. Cleaning that up could be a lot of trouble, though probably not quite as much as drawing the townships manually. Personally, I have enough TIGER data to fix; doing this the long way sounds like a nice change of pace. —Preceding unsigned comment added by Vid the Kid (talkcontribs) 10:34, 22 October 2009 (UTC)
It'd be nice enough just to get a textual definition of each township's boundaries. Does such a thing exist (appropriately licensed, of course)? – Minh Nguyễn (talk, contribs) 07:04, 25 October 2009 (UTC)
I don't know about a textual definition, but I might consider downloading TIGER '09 shapefiles and viewing them in ESRI's free ArcExplorer program. Using the County Subdivisions layer will show the townships and municipalities as areas, and the Edges layer will show pretty much everything as lines, so you can use the Identify tool on that layer to find out what roads or other features the township lines follow. In the case of Franklin County, the biggest hassle will be finding all of the little pockets of unincorporated areas scattered around Columbus. (Oddly enough, the TIGER '09 County Subdivisions layer for Franklin County omits the city of Hilliard, instead considering it all as Norwich Township...) More trouble than an import? Probably, but I think the work would be more satisfying. That said, I don't think I'll be doing any of this in the near future, as I have other priorities. Vid the Kid 09:25, 27 October 2009 (UTC)

An old question and a new question

For an answer to your question at User talk:Nakor/Interstate relations check, see Talk:Waychains TIGER fixup#Other useful tool.

I just read your "Primary, Secondary, Tertiary" essay, and I did things in a similar way in central Florida. Does this essentially match your vision? --NE2 03:54, 27 February 2010 (UTC)

That's a little bit denser with primary and secondary than I would probably do. But then, getting around Orlando is probably a lot different from getting around Columbus. I think in Florida cities, drivers rely on surface roads for longer distances than in Ohio cities. That will account for some of the extra secondary density. At least it's coherent, anyway. Actually, that seems to approximately match the density of primary and secondary roads found in European cities on OSM, so I don't think I can make much of a case for significant downgrades.

Question about

Around here, the county is inconsistent about abbreviations (on MUTCD-standard signage). Parkway may be Pkwy or Py; Boulevard may be Blvd or Bv. The shorter form usually appears on newer "next signal" green signs, but may also be on street signs mounted on traffic light poles. How would you handle this? --NE2 13:59, 14 April 2010 (UTC)

In that situation, I'd say the "consistently" part of my "commonly and consistently" criterion for using abbreviations is not met. Using the full form of the name suffixes might be the better way to go there. Vid the Kid 11:45, 25 April 2010 (UTC)
Then it's just silly, abbreviating some street names but not others. The worst part is that (if my memory is correct) they just recently started using the shorter forms; using your method, you'd have tagged them as Pkwy/Blvd back then, but then expanded them all once the county started with the shorter forms? This intersection is even worse - traffic light pole-mounted signs say "Turkey Lake Rd", "Palm Parkway", and "Central Fla. PKY", while the ground-mounted signs say "Central Florida PY" and (I believe) "Palm PY", and the advance guide signs say "Central Florida Py.", "Palm Py.", and "Turkey Lake Rd.". Get on I-4 right next door and it's "Central Fl. Pkwy." :) --NE2 07:40, 26 April 2010 (UTC)

Anyway, that argument is probably not winnable with the US OSM community in general. That's why one of the concepts I'd like to propose is a abbr_name=* key. Vid the Kid 11:45, 25 April 2010 (UTC)

Or perhaps a more generalized short_name=*. Either would be useful to renderers for squeezing street names into really small places, where the abbreviated form isn't especially obvious, such as "MLK Dr" for "Dr Martin Luther King Jr Dr". Same for Nominatim, which would have to deal with abbreviated input. – Minh Nguyễn (talk, contribs) 20:21, 25 April 2010 (UTC)


Not sure if you've seen Skobbler's bug map yet. Some of the bug reports are pretty helpful. – Minh Nguyễn (talk, contribs) 18:35, 9 November 2010 (UTC)

Township admin_levels

Shouldn't townships be admin_level=7? Villages and cities are part of townships until they withdraw from them. So far I've only been mapping townships in Hamilton County, where the suburbs have all withdrawn from their townships. But eventually I'll encounter a village or city that hasn't withdrawn, and I'd rather have the township be admin_level=7 then have the city admin_level=9 or 10, which would keep it from all but the highest zoom levels. (I'd also want to promote cities that have withdrawn to the same admin_level as the neighboring townships.) – Minh Nguyễn (talk, contribs) 19:21, 17 October 2012 (BST)

It is my understanding that all incorporated cities are "withdrawn" from townships. As townships and incorporated cities never overlap, and the entire state of Ohio (except perhaps for Lake Erie) is contiguously covered by townships and incorporated cities, they are essentially equivalent units of government. Am I wrong about that? The Census Bureau seems to make some kind of distinction between Incorporated Places (which are still parts of townships) and Independent Cites (which are not) but I thought that was just more statistical BS like CDPs. Census says Hilliard is an Incorporated Place, part of (mostly) Norwich Township, but my local experience suggests otherwise; Norwich Township Fire Dept does indeed serve Hilliard but that is per an explicit contract (and they also serve Brown Township per another contract). I'd like to see an authoritative source describing how this works in Ohio. If it turns out some cities are parts of townships and some aren't, then I think maybe the admin_levels should look something like:
  • County: 6 (as previously established)
  • Independent City: 7
  • Township: 8
  • Incorporated Place: 9
  • City-designated subdivisions such as wards: 10
  • Census-Designated Place: no administrative boundary at all (as previously established)
But I would really prefer not to be wrong :-) Vid the Kid 23:04, 2 November 2012 (UTC)

I'm not sure if it's unique to Ohio, but a city or village by default belongs to the township it was created out of. In order for a city to withdraw, it has to be coextensive with the township. However, it can petition the county board of commissioners to carve out a new paper township coextensive with the city under ORC 703.22, with the effect that the city alone has taxing power over its businesses and residents, who in turn can no longer vote for township trustees. Paper townships often share their city's name, but Louisville for instance is coextensive with "Constitution Township", after the city's nickname. It looks to me like the Census' concepts of "independent cities" and "legal county subdivisions" maps cleanly to withdrawn cities and townships, except for the fact that the Census assumes that an independent city's and its legal county subdivision share the same name. [4]

Lots of cities have withdrawn from their counties, but the distinction isn't theoretical. Page 5 of these county commissioners' minutes lists some examples in Medina County where villages haven't withdrawn. (Can they?) A fun example is the city of Loveland, which sits in three counties. On paper, there's a Loveland Township in each county, each corresponding to the city limits. That wasn't quite the case a few years ago, when Loveland annexed some land in Hamilton Township. Since the township boundaries don't automatically change as a result of annexation, the few residents on that land owed property taxes to both jurisdictions and could vote for both trustees and city council. Loveland had to petition the Warren County commissioners to synchronize the boundaries and kick the residents out of Hamilton Township.

It's hard to find current examples of cities that haven't withdrawn without simply referring to Census maps. But Fairfield's Heritage Township was only created in 1995, 40 years after the city was incorporated. The Hamilton County's resolution and legal opinion sheds some light on the process.

Regardless, I'm starting to think it would make sense for withdrawn cities to have the same admin_level as townships, and everything else a higher admin_level. The whole point to withdrawing is that cities can become peers with the townships. Paper township names that don't match their cities', like "Constitution Township", can go in alt_name. Where the paper township boundaries aren't up to date, the city and actual township boundaries can overlap. How's that sound?

 – Minh Nguyễn (talk, contribs) 08:22, 3 November 2012 (UTC)

Reality is being a pain in my ass. I still don't trust the Census maps regarding which cities have "withdrawn". (It also says that withdrawn cities, what it calls "independent cities," aren't even part of any county, IIRC.) But I've gotten used to the idea that some incorporated cities do in fact overlap township(s). I'm also thinking that admin_level=8 is appropriate for all incorporated cities (even those that have not withdrawn) as well as all townships. For if a city is incorporated, then I would indeed consider it a peer with townships, especially the township(s) in which it exists.
I'm also thinking that, if making a pretty map, township boundaries that are within a non-withdrawn incorporated city should be drawn less prominently. Pre-processing can detect this but it would be a royal pain to code. Might there be a sensible way to tag those boundaries differently (maybe with a different value for border_type) that's not blatantly tagging for the renderer? Vid the Kid 18:36, 12 January 2013 (UTC)
The Census maps are out of date, at least in Southwest Ohio, because there has been a strong push in the last few years for all the area's cities and villages to withdraw, to avoid the "possibility of double taxation". The concept of an "independent city" independent of a county is intended primarily for Virginia and a few edge cases around the country. Ignore that bit.
Based on a cursory reading of ORC 503.07–09, it still seems to me that a non-withdrawn municipality is considered a part of one or more townships, even though the townships may provide fewer services directly (except perhaps in the case of a limited home rule township). I only know of a few non-withdrawn cities in villages in SW Ohio, and while I would consider them basically peers with withdrawn cities, but not peers with their townships.
If we tag all townships and municipalities at admin_level=8, it just seems weird that a given location could be part of two different areas with the same admin_level. The map would be a series of Venn diagrams. Better, in my opinion, to have the scheme I detailed in United States admin level, in which townships are admin_level=7 (regardless of limited home rule) and all municipalities are admin_level=8 (regardless of withdrawal). Thus, some locations would lie inside admin_level=7 while others don't, reflecting the fact that some places have an additional township level of government while others don't. Sure, you'd end up with admin_level=7 township lines cutting through admin_level=8 non-withdrawn cities, but I'm inclined to think these situations are rare, because cities that straddle township lines are very likely to withdraw anyways. Besides, we already have admin_level=6 county lines cutting through admin_level=8 cities.
Every scheme has its pitfalls, because Ohio's system is so complex. Withdrawn cities should be peers with townships, but the scheme at United States admin level would make townships "first among equals" inside a county.
 – Minh Nguyễn (talk, contribs) 07:19, 13 January 2013 (UTC)
I can live with the 7&8 scheme you've described. Vid the Kid 06:13, 22 January 2013 (UTC)
Something just occurred to me. If withdrawn cities are conterminous with paper townships, then shouldn't they effectively get level 7 borders? Except of course for cases where a city has annexed land but the township borders haven't been changed, in which case the "new" city boundary would be level 8… Vid the Kid 19:15, 27 January 2013 (UTC)
That's what I proposed above, back in November, before changing my mind back to the all townships=7 / all municipalities=8 scheme. This whole time, I've been on the fence about withdrawn municipalities, because on the one hand they do have township+municipality authority, but on the other hand, most people can't tell the difference between a city that has withdrawn and one that hasn't. (Of course, someone who lives in a non-withdrawn city may know, because they still pay taxes to the township and vote for trustees, but might have friends in withdrawn cities who don't.) So either way works for me, honestly. If we go with admin_level=7 for withdrawn municipalities, then I just have a lot of work to do around Cincinnati. :^)
Now, if we're being pedantic, a withdrawn city should have two relations, one for the city proper and one for the paper township. Then, if the city annexes some land but the paper township isn't updated, the city relation would extend beyond the paper township's borders. That's consistent with Ohio law, but since OSM frowns upon fiction, especially legal fiction, a separate relation for the paper township seems wrong. Plus, I like the rule of thumb about mapping only the jurisdictions that get their own welcome signs.
So how about this: all townships are admin_level=7; non-withdrawn municipalities are admin_level=8; and withdrawn municipalities are admin_level=7, except when there are non-withdrawn portions, in which case we map the paper township – which ceases to be defunct, but don't tell the county commissioners, lest they hold an election for "paper township trustees" (!) – at admin_level=7 and the municipality at admin_level=8. So if Loveland again annexes land and forgets to annex it to Loveland Township, we'd have three Loveland Township relations at admin_level=7 (one per county) and one Loveland relation at admin_level=8. Messy, but I could live with it.
Like I said, every scheme has its pitfalls. February 1st is Ohio Township Day, in which schoolchildren learn about Ohio's arcane jurisdictional laws. Glad we're getting a head start. ;^) I just finished mapping Hamilton County's gerrymandered townships (at admin_level=7) and hope to contact some of them soon about verifying my work.
 – Minh Nguyễn (talk, contribs) 08:27, 28 January 2013 (UTC)