OpenStreetMap

Hmm, I'm seeing three places in the same forest all called Gemeindewald. That seems duplicate to me, considering this best practice: https://wiki.openstreetmap.org/wiki/One_feature,_one_OSM_element
Has this issue been discussed in the (german) osm community somewhere?

I've also got doubts about adding places allegedly called "at the little street". This is not at all unique for a locality (https://www.openstreetmap.org/search?query=Am%20Str%C3%A4%C3%9Fchen) and there are also many real streets called as such (e.g. https://www.openstreetmap.org/way/186657234). This will just muddle regular address searches and I wonder if anyone really calls a locality such a name. How do you see this?

I just noticed another questionable locality name: Glücksburgweg (https://www.openstreetmap.org/node/9805658383). That sounds like a street name to me, and indeed there is a street "Glücksburg" just to the north. Do people really call this field (that you placed the node into), or the greater forest or Camp Astrid area, "Glücksburgweg"?

Hi LucGommans,

thank you for your comments. The data I add is from the official ALKIS database, which contains all that names. In German these points are called "Gewanne". To map those as place/locality is described in the German OSM wiki for place/locality.

The names are shown in nearly exactly the same way as they appear in official digital topographical maps from Germany (depending on scale).

The repetition in some cases is due to the large area of that object. It is comparable to a long street which also uses the street name several times, not only once.

And yes, the addition of "Gewanne" has been discussed many time in the German OSM forum.

Best Regards,
physi

Hi,

"And yes, the addition of "Gewanne" has been discussed many time in the German OSM forum."

Indeed, it has been discussed a couple of times with the question of how to map and if to map at all:
https://forum.openstreetmap.org/viewtopic.php?id=71409
https://forum.openstreetmap.org/viewtopic.php?id=75785

Both topics with **no progress in therms of a community accepted approach** due to a lack of participation of the change originator.

BR
tux67

Thanks physi for the response and thanks Tux for the links!

--- Tagging scheme ---

I see the community is quite divided about whether these are useful, and even less sure about whether place=locality is a good tag. Adding these locality=ALKIS_Lage tags until we found a good tagging scheme, and then mass editing them later, sounds okay to me as a tagging scheme.

Of course, this assumes that you then also make sure a suitable scheme is found (because you are the one that wants to add this), and clean up the temporary tags afterwards. Otherwise the temporary tags will just stay there forever. So far, I am not seeing a lot of participation to make this happen, but I do see you are continuing to add nodes on the map for more than a year since the discussions started.

--- Currentness ---

I see there is a post from physi in one of the threads, saying:

"Inhaltlich bin ich der Ansicht, dass die Eintragung einer historischen Flurstück- oder Gewannbezeichnung dann sinnvollerweise erfolgt, wenn der Name noch gebräuchlich oder bekannt ist. Dazu sind aber sehr lokale Kenntnisse notwendig."

This makes sense to me, so then we should stop copying these names from a database and instead use local knowledge, right?

The ALKIS database might contain these labels in some "current" layer because it is the latest info it has available, but that doesn't mean people on the ground call it by this name anymore. It just means that ALKIS users might use the name to refer to the place. Copying labels from a database like this, without any knowledge of local use, seems similar to tagging what3words labels as localities. Maybe someone uses it, maybe not.

In another changeset discussion (https://www.openstreetmap.org/changeset/122333151) you wrote "Frage der Gebräuchlichkeit halte ich dagegen für relevant bei der Darstellung dieser Informationen". This is not true: the renderer can't choose that old names are irrelevant, because it's not labeled as an old name. If a place is only formerly known by a name, then the name=* tag is simply not accurate. For example Westerwald (https://www.openstreetmap.org/node/9815156407) placed outside of a Wald... whether that Feld is still called Wald, is something only locals can say.

--- Accuracy ---

You haven't commented on the issue I raised above, where Camp Astrid allegedly is a locality called Glücksburgweg by the locals, while in reality there is a real Glücksburgsweg (and bus stop called Glücksburg) only a few minutes walking to the north. Not being from Germany originally, I cannot say whether the names near me are accurate, but this is one example where it seems clear that this ALKIS data placed a label in the wrong place.

I don't know what the threshold is before OSM considers a data source reliable and suitable to copy from (even if there are, of course, occasional mistakes). It might be good to investigate ALKIS' currentness and accuracy before using it.

--- Specificity ---

Labels such as "An dem Rothenbach" (e.g. https://www.openstreetmap.org/node/9816379762) obviously refer to an elongated area, namely the length of the brook called Rothenbach. Adding a locality here rather than a name or old_name to the brook would be more correct and help users a lot more to find something that is described as being "by some brook". Picking a random point and calling that place "by the brook" is not very specific. Since the brook currently has no name, if this is the brook that ALKIS refers to, then adding the name there would be a useful addition to the map!

Similarly for the Eschweiler forest, where the Stadtwald is allegedly also called Gemeindewald, it would make more sense to add the loc_name to the existing Wald object (https://www.openstreetmap.org/relation/106123) rather than adding a duplicate object to a few locations that are seemingly chosen at random (or for/by the ALKIS renderer). I would see this as a useful addition!

--- Duplication ---

"The repetition in some cases is due to the large area of that object. It is comparable to a long street which also uses the street name several times, not only once."

That's because they indicate a defined region, from point X until point Y this road is called Hauptstrasse, and for technical reasons the way object needs to be split into multiple segments (to have different tags for different parts). This is not the same as adding duplicate localities for the same locality.

If there are adjacent, duplicate locality names shown in ALKIS data, and it is known to be current and accurate, then one should add a single node at a suitable central point, rather than adding multiple nodes for the same place.

(Update: while I was writing this, I see you added more changesets, and this one for example with even more duplication than I saw in past changesets... https://www.openstreetmap.org/changeset/122553475 Please stop that, I really don't think this is correct.)

--- Community ---

Working together always requires communication and discussing with others.

Aixbrick requested 10 days ago: "Bitte höre vorerst auf, Gewannnamen zu ergänzen." (https://www.openstreetmap.org/changeset/122076044)

Vademecum said: "Sich der Beteiligung an der Diskussion zu verweigern, wenn es um einen von dir erfundenen strittigen Value geht, finde ich im Rahmen eines Gemeinschaftsprojekts bedauerlich." (same changeset)

flohoff wrote 3 days ago: "Lagebezeichnungen zu übernehmen die typischerweise ausserhalb des Katasters niemand benutzt macht IMHO keinen Sinn." but asked for discussion (https://forum.openstreetmap.org/viewtopic.php?id=75785)

I think I also raised a few points that should give pause regarding adding more of these names from ALKIS.

The OSM community can be conservative (particularly regarding imports) and we have to be weary of ossification or excessive bureaucracy. In this case, multiple people question the source (whether the data is good at all) and everyone seems to agree that the tagging scheme isn't good either. The latter is okay temporarily, if it can be machine-edited later, but more than a year is not very temporary.

Because of the questionable source quality, combined with repeated requests from different people:

you should stop adding more of these names, at least until the data source's quality is known (then we can still see about the tagging scheme and mechanically edit it).

Just 4 points:

1. The ALKIS data is by no means a "questionable datasource". On the contrary, it is the official ("amtliche") datasource for whole Germany since 2015. I would recommend to acquaint yourself with ALKIS.
2. Many thousands of data points from the ALKIS system (or derived maps from it) have been entered into OSM in the last years, and not by me. This was (and still is) in accordance with the German wiki for place/locality.
3. In the past, I myself tried to "hide" some of these points by changing the tagging to "old_name". Then you very quickly will find a group of people who strongly oppose to hiding these data. And if you add more data instead, you will find another group of people opposing to that. By the way, both are actually quite small groups. So only if you do nothing both groups are happy...
4.
Glücksburgweg is correct, check the ALKIS data yourself (geoportal.nrw)

1 & 4: I don't doubt that ALKIS is official from the government (that doesn't make it correct) or that you managed to use Strg+C correctly (instead, I highly doubt that someone would call the place Glücksburgweg when there is a Weg called Glücksburg at walking distance to the north).

2: okay so others made the same mistake, is that a reason to continue?

3: not all points are historic, I'm sure many of them are also current, but you said yourself (and I agree) that, without local knowledge, it cannot be said which belong in which category.

Is it possible that you don't understand what kind of names we are discussing here?
Those are "Flurnamen" (or "Gewanne"). And the important point is that to each (!) "Flurstück" there is a name. There is not a single point in all of Germany that does not have a name.
In case of populated areas the "Flurnamen" typically are the street names plus the house number. And in the same way, street names may have evolved over time but are now officially named, this is true for non populated areas. ALKIS sets the official name. In this sense it can not be wrong. (There may be errors as in any big database of course)

Just a clarification to what I wrote before:
I said: To each (!) "Flurstück" there is a name.
It would be more correct to say:
Each (!) "Flurstück" is part of a "Lage", which in turn has a name as defined in ALKIS. These "Lagen" may contain for example "Fluren" or "Gewanne".