I would like to bring up a small point of discussion from a recent changeset here in the forum.
The changeset edited a route relation whose name already contains the start and end locations (“Jakobsweg Brochenzell - Nonnenhorn”). In the changeset, the from and to tags were removed, with the argument that they contained exactly the same place names as the name.
When asked, the reason for the change was that in various map overlays and other tools, from and to are displayed in addition to the name, which duplicates the place names. This seems unnecessarily redundant in some places.
But I could also imagine that this could possibly be interpreted as “tagging for the renderer” – and that is precisely why I think it makes sense to clarify this here in the forum. So far, I have not found any clear recommendation in the wiki on how to deal with from/to in routes if this information is already included in the name. So the question is:
Is there an established consensus on whether from/to are redundant and dispensable in such cases, or should they always be set, even if they only duplicate the name?
If anything the practice of adding descriptive names should be abandoned.
Removing valid tags just because they can be derived from descriptive names is terrible, should be stopped and reverted.
Just because some tool properly uses from and to should not be a reason to remove them!
And yes, I know that there is long standing terrible advise at wiki encouraging to add fake names.
Ideally we would change it and use name tag for names, also for relations. And purge mistagged name keys. This would enable for example Nominatim to enable search by names for relations.
Agreed. Usually all of the numerous segments of the “Jakobsweg” (also called Jacobusweg) have exactly this name, whereas the " … Brochenzell-Nonnenhorn" is just an addition of the from-to locations.
I would change the name to “Jakobsweg” only and leave the from and to tags untouched.
The reason for the introduction of from and to (and via by the way) was to avoid constructing descriptive names. In nameshould go only the name, and any descriptive information should go into dedicated tags, this is common standard.
hmmm, maybe examine the PT naming policy/practice in Italy… mode, nr, from → to all in the route name (but I’ve been avoiding PT work like the bubolics, only done the brand new trolleybus and gap closing in recent years on routes I’ve touched in past. Think the destination of whats shown on the front or side of the PT vehicles is OK, from where it came truly superfluous in the route name, IMO.
I do not see any reason why the hundreds of route segments of the Jakobsweg should not have simply the name “Jakobsweg”, same as the superroute and the (probably missing) global hyperroute. “From-to” or other descriptive parts like “Part 1” should not be included in the name imo.
If you talk to people hiking there they’ll tell you they are hiking the “Jakobsweg”, not "Jakobsweg Bremen-Fulda-Herbstein (Rhön) Part 1". Such details can be part of the description, where in many cases the descriptive name is repeated once again redundantly anyhow.
I would like to point out that it’s entirely OK if a route does not have a name, as well. Routes I would give noname=yes to would be every MAX line in Portland (along with noref=yes and the appropriate colour=* tag) and most route=road relations in the US (it’s borderline rare for routes to be named in the US, usually limited to landmark routes; some notable exceptions would be tourism routes (like the Oklahoma Fishing Trail or the Oregon Outback Scenic Byway) and pre-route-numbering historic routes (like the Lincoln Highway).
The whole route has a name, for sure, not really a small part as here.
Parts in Spain are named:
Camino Francés: Saint-Jean-Pied-de-Port → Santiago
Camino Aragonés
Camino del Norte: Irún → Santiago
Camino de la Plata: Séville → Astorga
Camino Sanabrés: Granja de la Moreruela (Zamora) → Santiago
Camino de Levante: Valence → Zamora
Camino de Madrid: Madrid → Sahagún
Camino del Ebro: Sant Jaume d’Enveja → Logroño
Camino Mozarabe: Grenade → Mérida
Camino Interno: Irún → Santo Domingo de la Calzada
By now, enough forum and mailing list threads have come to the same conclusion against this constructed name format (ASCII art or no) that if someone were to unceremoniously mark it as deprecated on the wiki, the great collective sigh of relief would likely drown out any staunch opposition based on principle. The only difficulty would be retagging existing relations without creating giant changeset bboxes.
In the meantime, local communities have been abandoning the PTv2 name format one by one based on local real-world conditions. This process can start in any local community independently of what the wiki says.
Those dealing with correctly named route relations would probably benefit from JOSM, the website, and other tools constructing these constructed labels when listing or linking to relations. The label can omit redundant details if the name contains an arrow of any sort, so that mappers don’t perceive a need to delete the more structured tags.
The passage in question has been marked as controversial for years, so it’s not like this comes as much of a surprise to those who read the wiki in isolation. To the contrary, the downsides of this format are well-recognized, and the dataloss that spurred this thread demonstrates the need for more decisive action.
Due to its size and flaws, the PTv2 schema has proven more malleable than many approved proposals. For example, public_transport:version=2 became optional, and → was introduced as an alternative to =>. Both changes happened after much less discussion than we’ve had about dropping the naming formula. I don’t think there’s much risk that further tweaking this small part of the scheme once again would severely undermine the legitimacy of the proposal process.
The more important consideration is whether JOSM users are prepared to deal with identically labeled relations in the Relations pane. That’s the one consumer of this formula that depends on it, so to speak.
In my opinion OSM wiki should be the de-facto place where people can learn how to properly map. It shouldn’t be “this and this, but ignore that”. Not everybody should, can or want to spend hours on Discord/Slack/this forum to sift through for exclusions and corrections to the written documentation.
And aren’t “wiki editors” just us, OSM editors? Why have such separation?
