Hei!

Ser du har satt access=no på Jomfrurunden. Er det noe spesiell grunn til det?

Just noticed the large forest north of Dale being called Norafjellhalvøya in this changeset, and I found it weird for two reasons:
1. It is a name that should normally be assigned to a peninsula, rather than a wood.
2. I can't find any official sources mentioning this name.

Would you mind sharing your source of this name?

I am getting the feeling you're not a big fan, no. I'd love to hear you out on that, though. I know you have far more experience than me with regards to mapping these structures, hence, if you have any thoughts on why you prefer areas over the site relation in this case, in addition to the things you've mentioned so far, please let me know. I want to learn and understand.

> If you want to keep doing it like this I can't stop you, but I'll hand over power mapping.
I've got little interest in mapping power infrastructure in general. I got into it because I'm somewhat completionist with regards to Osterøy, due to personal ties. I've already covered the four power plants on the island, meaning I'm probably happy for a long time.

> I can't be part of a mapping practice I don't agree with.
Could you explain what you mean by "being part of" this? I hope I don't come off as wanting to force my mapping practice on you. This is my preferred style for now, but that doesn't mean it needs to be yours as well.

> There are around 900-1000 hydro power stations still missing that I had planned to map.
I suggest you go for it, just like you've always had. In general, I find it a vast improvement that these stations are mapped at all, even if they aren't in the style I'd choose myself. As I said, I have no interest in pushing this style on you.

Not sure I recognise the vastness of these plants as a problem. That seems to apply to wind farms as well. See for instance this one: relation/11343101#map=11/53.9769/1.7080 In fact, it seems this quality of being spread out over a large area is exactly why you might want to use a site relation, rather than an area. At least from my understanding of the site relation. I recognise that the _rendering_ as a large, odd-shaped multi-polygon is problematic, but this is a _choice_ made by the data consumer, and I don't want to adapt my mapping style only to please a renderer, if I feel I lose accuracy in the process. I'll get back to what I mean by that.

As for consensus, I can't say there seems to be a consensus against it either. This approved proposal describes mapping hydro power as site relations, for instance, which seems to contradict what you are saying regarding voting: osm.wiki/Proposed_features/Power_generation_refinement#Dispersed_facilities_power_plants

> As long as there is no defined center point for a site relation on hydro plants, using relation will cause a big mess and a very ugly map.

I disagree with this. In fact, I think the data consumer benefits from having the freedom and flexibility to decide on where to put the center point themselves, rather than having the mapper do it for them. There is an analogy here to how we, as mappers, don't decide where or how the name of a lake is rendered. Instead, different renderers use different approaches to decide on this, depending on their priorities. This  will cause a big mess and a very ugly map, only if the consumer makes poor choices with the data they are given access to. If we want better rendering, we should direct our efforts at helping renderers make better choices with the data at hand.

> If a consumer is using generators as center point, what happens if generators are missing?
Then you could fall back to something else, like the substation connecting the plant to the grid, or perhaps the dam(s), depending on what makes more sense. Finally you could fall back to the full plant center point, if nothing else makes sense, but that would indeed be one weird power plant.

> Unless there is a proper system for site relation with roles aimed at power plants for each component it is going to be a mess for them to figure out the center point.
I don't think that is needed. In many ways the elements themselves contain the information you need. Something tagged as being a dam, or a reservoir, and is part of a hydro plant, is presumably storing water for said plant. Similarly, headraces or tailraces are usually tagged with a usage tag. Unless a category of members can play multiple roles, and this isn't tagged directly on the members,  I'm not sure specifying the role is needed.

So, in summary, I recognise that this power plant looks ugly on openinframap right now. I sympathise with wanting it to look differently, and do so myself. In that regard, I am your ally.

I just disagree with you that the database should change to facilitate that, and want the renderer to improve instead. I have explained why I think the renderer is making poor choices. Now, let me explain why I actually prefer mapping with the site relation, and think it makes for more accurate data.

Reason 1: osm.wiki/One_feature,_one_OSM_element
- I consider dams, reservoirs, transfer tunnels and penstocks as part of a hydro power plant, and I don't think that is controversial. These are man-made structures with the sole purpose of facilitating power generation at the plant's generator(s).- I can represent this on the map by either drawing a really big area, if there is a perimeter, but in this case there is not, so the alternative is then a site relation.- So, in order to map this power plant, with all its parts, I have no choice but to do what I have done. IMO drawing a smaller area around only the generators is actually lying about the full extent of the infrastructure needed to generate the power.

Reason 2: Giving context to other mappers.
When coming across a dam, or a reservoir, either in real life, or on map, I can often find myself asking, what role does this play in a larger system? The benefit of the mapping style I've used here becomes apparent when looking at these items in an editor. This reservoir: relation/4833629 now gives me a link to the larger power plant relation: relation/13697166#map=13/60.5680/5.6134, which means I can understand what is going on at a system level so much quicker.

These are the benefits I don't want to lose in order to help the renderer out. I hope you can also see why I think this is valuable.

I'm not sure what I would advise the next steps to be. I was going to suggest creating an issue with openinframap, but I see you've already done so: https://github.com/openinframap/openinframap/issues/20, without being met with much response. That makes me sad.

I see your point. I noticed myself that the rendering in OpenInfraMap ended up looking ugly. Both in terms of center point and surrounding polygon. Nonetheless, I decided not to revert this mapping style, because I feel like it’s the renderer failing rather than the mapping being inaccurate. And I don’t want to be tagging for the renderer. Indeed, reading the Tagging Guidelines section of power=plant, it seems the mapping practice I’ve used is fine, or even recommended.

I guess it could be argued that openinframap should use the center point of the generators in a site, rather than the center point of the bounding polygon.

No worries! That can happen to the best of us.

For context: This way featured in a discussion about "fuzzy areas" on the tagging mailing list. I wanted to have a look since it was mentioned and found it deleted.

This changeset deleted this way: way/211234843

That seems a rather controversial move.
Want to elaborate on reasons?

Hi,

I didn't know about said tags, but one of them was a good fit, so I've adopted it.

Was trying to refer to to the id in this database entry: https://www.nve.no/energi/energisystem/vannkraft/vannkraftdatabase/vannkraftverk/?id=327

That's the ID used by the The Norwegian Water Resources and Energy Directorate.

Thanks! Might want to specify in the ticket that we specifically would like the "Missing power tower/pole within power line" warning to go away when two distinct lines are connected using `power=connection`. They might also want to reword the error message for connected lines without said tag, in order to also suggest this tag, instead of just tower or pole.

It doesn't resolve the JOSM warnings, but at least it is a statement that the mapping is as intended, and perhaps the warnings will be resolved this way in future updates.

I've been digging around and found this: osm.wiki/Tag%3Apower%3Dconnection

It only has status: proposed, but seems fairly uncontroversial, and seems like an exact match for what we're after.

I tagged the nodes as suggested in the wiki in this changeset: changeset/113828378

I learnt something new today! Thank you for this thorough and high quality feedback.

To me, it seems like it would be natural to have a tag for this; something like `power=tap`. There is no similar tag, as far as I can see. Having one would at least help remove a false positive warning in this case, and probably quite a few more, since this is fairly common. What do you think about it? Have you heard of such a proposal being reviewed before?

What is the thinking behind this reversion?

The reason I modified this junction in the first place was because JOSM complained about power line noes that weren't towers. I agree with JOSM: I don't see it as likely that these power lines are connected in mid air, as opposed to a connection through the nearby towers.

If you have a reason to believe the warning is wrong, I suggest adding a note to the untagged node connection explaining why it is as intended.

Am done mending most trivial issues resulting from the Vik import, btw. So if you have any feeling of ownership to the area, please have a look and see if there is anything you want to address, post import.

The import has taken quite a bit of effort, so will take a break for the next couple of weeks, but I know there are quite a few nice trails in the south of the county (Stølsheimen) in need of refinement. Might have a look at that again in a while if nobody else gets to it.

> The resolution it's currently showing we are not legally allowed to use. Keep that in mind.

Fair point, and I'll keep that in mind until the layer is mended. Although in this particular case I assume there isn't any difference between this source and other sources we would be allowed to use, hence it's hard to argue anything was gained from using it.

Thanks for the general advice. I appreciate it.

Thank you for catching this! The changes were based on outdated aerials, as you seem to have suspected.

I have now corrected the road based on N50 (the iD layer, which seems to be a lot more detailed than N50). It seems to look more or less like I remember it from before I touched it.

Feel free to review it and tell me what you think.

Sorry about the sloppy work.

I found an English translation of the
Outdoor Recreation Act (or Friluftslov) that both me and the complaint is referring to. The translation is not considered official, but is written by the government, and more or less conveys the information content of the Norwegian law: https://www.regjeringen.no/en/dokumenter/outdoor-recreation-act/id172932/

I hope it can help is clearing up questions you might have. Of most relevance are §2, §3, and §4.

CONTINUED:

My reading of the complaint is that they take issue with some section of path, or road, that crosses "innmark" (cultivated land), as I can see no conceivable argument as to why access should be restricted on the parts of these trails that cross uncultivated land. Hence, I assume the path you point out in your initial comment is not the path segment the complainant has issue with.

It would be beneficial if the complainant could specify the exact path segments that would fall under one of the three reasons listed above for possibly limiting access, as well as their reasoning in accordance with the law. Only then would it be practical to evaluate whether their complaint has merit.

Given a complaint has been issued, I feel obliged to help you resolve the matter to the best of my abilities. Although it is not clear which specific trail the complainant is referring to, I am responsible for most trails on the mountain, so I am probably the person to get in touch with anyway. I am willing to keep up the dialogue here on this changeset, or we can move to a better suited platform if you wish.

First, a bit about my background. I would consider myself semi-local to the area. I grew up in the nearby village of Valestrandsfossen (to the south), and have walked most of these paths myself. I haven't lived there for the last ten years, but I have family and friends there, so I visit multiple times a year.

Now, addressing the claims in the complaint. They seem to me to be an inaccurate representation of Norwegian law, to the point where it is unclear whether the complainant has a point or not. Roughly speaking, the law in question divides land into two categories:

Utmark (roughly translated: "Uncultivated land")
In law, this is defined as "everything that isn't considered cultivated land", but more or less covers marshland, forests, shrub land, mountains and the like. This category of land comes with an extensive right to roam, including camping for shorter periods of time, as long as you leave no trace and generally behave yourself. I could go into more detail, but this doesn't seem relevant to the case.

Innmark (roughly translated: "Cultivated land")
This category is defined as courtyards, gardens, cultivated fields, industrial areas, and similar. That is, land with a significant human footprint. This is the category of land I assume that the complainant is referring to when mentioning "farms".
This category offers less access rights than the previous category, e.g. camping on cultivated land is forbidden unless permission is obtained from the land owner. But unlike what the complaint is implying, there are still extensive access rights to consider for cultivated land, which I will cover.

The law deals with two different forms of access to cultivated land; on road and off road access. Off road access (e.g. crossing crop fields outside a clear path or road), is allowed in the period 15th of october to 29th of april, as long as the ground is frozen and/or covered in snow. On road access over private land is also allowed, all year around, under certain conditions. The owner is allowed to ban motorised traffic on private roads, but access by foot, bicycle or horse is explicitly allowed. The owner is also allowed to ban access by bicycle or horse on more "informal" paths, as seems to be the case here, but even on informal paths you have unconditional access rights, as long as you travel by foot. (Hence the argument that "These paths are not marked and not in official maps" is not relevant.)

I say unconditional, but that isn't entirely true. There are still scenarios that might allow the owner to block access, but it is unclear from the complaint whether those scenarios are relevant here. I will summarise them here, but the details are available at this link https://lovdata.no/lov/1957-06-28-16/§3a:

1. Passage can be banned on roads or paths that go through someone's garden or courtyard (i.e. where it would cause significant nuisance). So, for instance, a stereotypical driveway would not imply access rights, but a private road that does not pass though someone's garden would not be problematic. In most cases, though, this shouldn't be too hard to judge from aerial imagery.
2. The municipality can issue a ban on access.
3. The owner can ban access if access would result in "significant damage".

The complainant would have to make a strong case for one of the exceptions above, which they don't seem to be doing. All the information I summarised above is available in the link I provided, so it would make sense to reference that.

woodpeck, for uncultivated land specifically, right to roam is given by law in Norway: https://lovdata.no/lov/1957-06-28-16/§2

Hence, whether or not this is private property isn't a concern, as long as you show appropriate care and respect for said property.