Sanderd17 has commented on the following diary entries

Post When Comment
Starting around my local area about 13 hours ago

Hi winnieryan,

The original data in the USA mostly comes from the free governmental TIGER dataset. This dataset had a basic road layout, but overall a low quality with lots of problems.

Other data (churches, post offices, ...) are normally added by volunteers (like you now).

If everyone maps his own house (or street), the entire world will be mapped.

PS. If you want to see what an active mapping community can do, take a look at Germany (Bremen f.e.: ), in Germany, governmental data isn't free, so everything you see is made by volunteers. I hope that a look at a well-mapped place inspires you to see what can be done.

Mapping Andhra University College of Engineering (Andhra Pradesh, India) using Field Papers: 15 days ago

Something not yet mentioned: you have some ways with only a name tag. F.e.

This won't show up in any renderer or other data user, instead you should try to describe the feature with tags (f.e. in this case).

And as mentioned by MarkusHD, names are for named features, so the name should be removed from that way.

Then, there are of course the buildings that should have straight corners (as mentioned before), but you also seem to follow building roofs. When imagery is taken at a non-straight angle (which depends on where the plane actually flew), then the imagery is georeferenced against the ground height. So only the ground floor of the building will be more or less correct, and the roof won't.

That said, I'm glad you started working on a local piece of map. That's the ideal way to learn mapping. When you have perfected this piece of ground, you will know a lot more to start doing typical mapbox work.

As you can't see the complete ground level, the best method is to draw around the roof, and then drag the building so it matches the ground floor.

Idea of issue tracker for tagging scheme 25 days ago

There's a big difference between software bugs (and even feature requests) and OSM tags.

Software is normally managed by a select group, they choose whether they want a feature or not, and whether a solution is good enough to solve a bug or implement a feature.

OSM tags on the other hand are more of an anarchy. Everyone can introduce new tags without any approval from others. In that anarchy, some tags grow naturally, and become a standard. Tools start using them, mappers see examples from other similar features, and the tag becomes more popular.

So introducing a new tag is, more than anything else, bootstrapping that positive-feedback loop. The well known wiki proposal process is just a part in bootstrapping that loop, by making the tag known and getting rid of any controversy around it.

So I don't believe bug reports will do anything except introducing yet another channel to discuss stuff and spread communication even more.

There are other things that could be done though, like helping tools to implement tags. Examples include: * creating a set of default tag implications. Like highway=residential inside the boundaries of Belgium implies maxspeed=50 * Making natural language translations of tags. Like admin_level=8 in Belgium can be translated to "Gemeente" in Dutch. * Adding categories to tags. Like amenity=pub belongs in the "food and drinks" category. * ...

By having data like that in a uniform format, tools can parse it directly, and it would help a lot with the creation of new tags IMO. By empowering tools, people will also start using these standardised tags more. And a project like this could be managed in a more centralised way, with a bug tracker and a team that decides which additions are good and which aren't.

Mapping private subdivision roads and other gated roads about 1 month ago

Where did you get the fee thing for access=permissive? AFAICS, access=permissive means it's owned by a private person or company, but the owner allows free traffic over the way (so no fee situation is no different from public roads, though they could theoretically as a toll, just like public roads).

In general, routers should know, if your start or end point is near a road with access=destination, access=private or any of those partial access tags, it's because you're already there, or you need to get there (and will take care of getting legal access). So the routers should be able to route you over those roads without problem, but they can't route you through.

When you tag the barrier, how would routers know if you can pass the barrier? There's no difference between a bollard to prevent through-traffic in a suburb road, or a bollard before a private road (f.e. an emergency road). Both could also be removable by authorised people (emergency operators, inhabitants, ...), so tagged with some other access tag than access=no.

I know that routers have difficulties with this (they usually just have passable or not passable road classes, and nothing in-between), but this isn't a reason to tag it differently. It's a reason to improve the routers out there.

Moreover, you're now making the access tags invisible (it's easy to render access tags on roads, while it's hard to render them on a barrier node), and the only reason it works now is because most routers still don't understand how access-tagged barriers work and just ignore them.

One extra argument: it's not because something is legally classified as having a certain access, that there's also a barrier enforcing that access. We have many streets that are marked as access=destination by a sign, but have no barrier whatsoever.

I do believe in a local tag creation process (we don't know how streets look like in the Philippines), but I think this way of tagging will seriously harm the OSM data consistency, and will make the job for routers a lot harder.

Natural language vs. abstract tags about 1 month ago

Very interesting post. I'm from Belgium, so not a native English speaker, but close enough to England to share a lot of the culture.

We personally have no problem with kiosks at all. But we do have problems with other tags, mostly the range between amenity=restaurant and amenity=pub.

In our language, a "café" is where you drink beer, so that's a pub. Then we have "taverne", "tearoom" and "brasserie", which is more to visit in the afternoon, so more like a "cafe" in OSM terms.

Distinguishing between our fast_food and restaurants is also hard. Certainly when it comes to our "fritures" ( ). A friture is a place where mostly fries and fried snacks are served. Ranging from take-away to having full service. However, they're not restaurants, as you need a license to have a restaurant title in Belgium. And IMO, it's very inappropriate to describe those fritures with full service as a fast_food amenity.

So we would like to introduce our custom tag amenity=friture, but nobody wants to tag objects like that, as it would remove all fritures from the visible map (while they are important, every Belgian village has at least one friture). And we can't get amenity=friture to render in mapnik before it gets used.

Some of the tags you mention also don't matter a lot. Like building=*, if you don't understand the tag, it's just a building, and doesn't matter which type. Most buildings are "yes" anyway.

10.000 broken Turn_Restriction in the OSM Planet File about 2 months ago

baditaflorin: a turn restriction can have more than 3 members. At least, it's possible to have multiple "via" members.

However, that type of turn restriction doesn't happen often in real life (it usually easier to set up physical barriers than signing the turn restriction), so I do think that most of your results are wrong, but there will be some correct turn restrictions there.

Using custom Mapbox layers for mapping missing features in JOSM about 2 months ago

Pratikyadav: IMO, it would be better to serve these layers on a browser site, with links to edit in JOSM, iD or P2.

Something similar to or . But of course, with the imagery background, other features can be highlighted.

shift in Village data about 2 months ago

There's no guarantee that the position of Bing images is correct to some degree.

So if you have no measured data (like GPS traces), and a lot of data has already been mapped to OSM, it's better to shift the images in order to match the data than the other way around.

I don't know what editor you use, but it's quite easy to add an offset to imagery in JOSM.

Private Plugin for "Faint" Trails? about 2 months ago

@Rovastar: visibility is of low importance when you have a map, and of even lower importance when you have a GPS device. As you can use the map or GPS device to locate you. That's the reason why not many maps render it. Contrary to sac_scale though, since that determines whether you'll be able to use the path at all.

@laridae, ok, but now you have some comments to convince your club ;)

As to finding a better renderer, if the area isn't too big, maybe you can display the tracks using Overpass API and the Overpass turbo styling ( ). See for an example (shows the paths mapped with surface=asphalt in black), you can use color codes, widths, dash patterns and opacity. Though it's currently not possible to stack different renderings with Overpass Turbo (which would be useful to get an outline color).

For the winter roads, see (or ice_road=*), these tags should also apply for paths, but I really doubt if any current renderer displays it.

Non-searchable = Non-existent about 2 months ago
  1. OSM is a data provider, not a service provider. So search isn't a service OSM should provide to the public, but providing it to help the mappers is enough.
  2. Google is a competitor, so it's normal they don't show our results anymore (they used to do in the past). When you look f.e. on DDG, you see that OSM appears on the firs page, while Google Maps doesn't:

Of course there are other search methods. My preference is the !osm bang of DDG (which is my default search engine). Note that I don't always like the search results from DDG, but it's so much easier to type !osm for OSM, !w for Wikipedia, !g for Google, !i for Google images, ... than having to go via links or selecting an alternative search engine.

Let's see. about 2 months ago

The most important difference between JOSM and other editors is that JOSM works with modes (by default).

So if you're in "add" mode (by pressing the 'a' hotkey), every click will add a point to a line (double-click to stop the line). If you are in selection mode (with the 's' hotkey), clicks will only select features (and allow you to tag them).

Then you have a lot of specialised modes (some need a plugin to use them), like building mode to draw rectangular buildings, or way-alignment mode where a click will move the points of a way.

Once you know that, learning JOSM gets easier.

Private Plugin for "Faint" Trails? about 2 months ago

@laridae, yes, the default renderer is not a specific renderer. When I mean a specialised renderer, I mean something like this: (see for alternatives)

Though that one also doesn't render the trail_visibility key, but does render the surface and sac_scale (it's limited to Europe because of server resources though).

The main map is not specialised at hiking, and users of the main map style should never expect something that renders as a path to be passable with sandals (though trail_visibility says nothing about the needed footwear, a path through a big grass field can also be more or less invisible, you should use sac_scale to show users what footwear they need to ware).

And it's true you would be giving the path data away. The whole point of OSM is about giving data away. I've worked many hours and days, just to give away the data I collected. According to recent stats, we've worked 61 labour years on the data, in a small country like Belgium:

But you can change your business model to be a "service" based one. As you see, there's a lack of good renderers for the data (and most of the good ones are limited to Europe), and even then, it's not handy to take your computer on a hiking course.

So you can contribute your data to OSM for free, and offer members a good rendering style to view the map (OSM data must be free, rendering styles can be protected by copyright separately and thus can be private). Or you could go even further and offer high-quality printouts to the members.

Private Plugin for "Faint" Trails? 2 months ago

This tag too:

Private Plugin for "Faint" Trails? 2 months ago

OSM data is public, if you want private data, you should use your own database (which can be combined with OSM data when you follow the license clausules closely).

But it's also possible to add those paths to the public database. There are certain tags you can use to mark paths as nearly invisible, or nearly impassable.

See * * * * *

When tagged correctly, dedicated maps will render it in a style so it's obvious for users.

Global Voting systems 2 months ago

Taking maps from Wikipedia will be best for you. OSM is too precise to download and render all boundary data easily. So it's better to use a Wikipedia map. You can take this map and update it with data you like (don't forget to upload it back again when you fixed the map). Updating isn't hard when you download it as an SVG file and open it in a program like Illustrator or Inkscape.

OSM doesn't gather data like the usage of voting systems either. So you won't find more info on that subject here.

Should we teach JOSM to first-time mapathon attendees? 2 months ago

Very interesting stats.

But next to the time mapped on the first event, it would probably also be nice to see the amount of work done on the mapathon (number of nodes, number of ways, number of tags, ...)

Measuring the number of elements isn't such a good way to determine the value of the work (it doesn't matter if all landuse is complete, when even the most important streets aren't present). But given that mapathons usually focus on a certain job, it should be more comparable than regular mapping I guess.

I agree with Richard that the lack of knowledge on iD is often a very real problem for potentional teachers. And I guess Joost would agree too, as the quest on the talk-be mailing list to find someone who feels comfortable with iD and wants to teach it is still going on.

Data and community in the Belgian regions 2 months ago

Great stats Joost, though I'm really scared by that one stat: 44% of nodes coming from one person in Flanders.

It's almost like having a monopoly on mapping. That's not really a healthy community. Who will maintain the work when he stops? Does he decide how things are mapped throughout the country?

It really frightens me.

Help with Franklin 2 months ago

Tbh, it's better to work locally. The original thought is: of everyone maps his own house, OSM will be a nap with incredible detail.

Now, that's of course an exaggeration, but I do find that everyone should stick to his own region (with the exception of humanitarian mapping, but that's just because they don't have the tools to map yet).

By sticking to your own region or city, you can add a lot more detail (like POI) that aren't visible on aerial pictures, out in any database. You're also more responsible of the area, so the data quality is higher and its more up-to-date.

If osm would just be a data dump, it would have been set up differently.

So I wish you all the best, but no, I won't help (you're doing great already btw).

5 years of Bing satellite imagery 3 months ago

Bing doesn't require a source=* tag. So it's practically impossible to figure out what impact Bing has.

I just know one thing: I've used Bing a lot while mapping, even when my source data was actually a survey. Thanks to Bing, the road network in Belgium was completed (it was impossible to find missing roads without aerial images).

Conflation engine Cygnus now in public beta 3 months ago

Are those cranes randomly generated on construction area?