I personally don’t like having “RIver” in the name tag, it feels like including a description in the name. Since people refer to rivers without the word “River”, I don’t think it’s part of the name. “This bridge goes across the Liffey” is a normal sentence and shows that “river” is not part of the name.
I’d like to remove “River” from name/name:en tags of rivers in Ireland. What do yous think?
Not in Ireland but I offer the New River as a counter-example. I mean I don’t think you’d want to name it as “New” on the map
Yes that one technically isn’t a river but it’s also not a canal which is what we have it as - it’s really more of an aqueduct. A quick search finds plenty of other instances of New River that are rivers though.
Two others to go with New River: Duke of Northumberland’s River, and Longford River (seems to lack a relation). Tagging of both of these might be harmonised with that of the New River.
I’m struggling to think of an Irish exception, but “Dutch River” (which OSM has as “River Don (Dutch River)”, complete with brackets) is another in England.
I think one thing to take into account is whether the name is shared with some other object. If I search for “Shannon” (the town), it might be useful to be able to clearly know that it isn’t the “River Shannon”. Given the number of re-used names in Ireland, this could be an issue.
Some rivers/streams/torrents have the specific inclusion in the name signs here what they are, mostly at bridges with a wavy water symbol below and e.g. 'Fiume Sangro". Quite a few of these signs revealed mistaggings of streams as rivers too and vice versa, barely a brook and tagged as river, some barely a brook yet officially qualified as river, but they look like it during storm rains.
That said, the actual official names are mostly without the qualification, it’s just only used in speech, like no one says River Rhine, it’s respectfully ‘The Rhine’.
Removing the ‘qualifiers’ in any form of automation would be wrong though, for who feels like it, manual checking which what I’ve done when one name appears in a mapping area moved into.
Isn’t this a fantastic argument for keeping “River” in name=*? How else would a data consumer know whether to append, prepend, or omit the word? Or is it truly always superfluous in Ireland, like saying “Dublin city”?
My argument is that “River” is not part of the name. That people can, and do, drop the {pre,suf}fix. The Wikipedia River Shannon article, quickly drops the prefix.
One could just as easily do a mechanical edit to switch all to prefix, or switch all to suffix, or remove all.
Many here have the prefix of ‘Fiume’ (River), ‘Torrente’ (stream) or ‘Fosso’ (literal translation a ditch, but very streaming, not like what the Dutch think of when they hear the word) Some prefixes on OSM are not official, but somehow since wikipedia has them in subject titles, it seems to be ‘Sir Wiki spoke’. The most famous here nearby ‘Torrente Piomba’, The Leaded Torrent. Never researched why it’s called that
I’m familiar with the practice of eliding generics – I do it myself all the time. Apparently the notion of the generic as part of the name has evolved over time, at least in some dialects of English:
In principle, a data consumer should be able to automatically elide the generic from name=*, leaving only the specific, to either avoid repetition or economize on available space. I’ve experimented with eliding generics for OSM Americana, but feedback was pretty ambivalent so I didn’t prioritize it at the time. On the other hand, I don’t know how I would reliably prepend or append “River” if the context calls for it, since the prefix and suffix aren’t normally interchangeable.
Anyhow, I wonder if you could even omit “River” without incurring dataloss. Over 4,100 “Rivers” in (the Republic of) Ireland are tagged waterway=stream rather than waterway=river. Not just any stream can be called a “River”.
First of, completely understand the logic and where you’re coming from on this as I’ve had a similar thought on this exact topic in the past. That being said, I’ve since changed my mind as the “River” part does form part of the name in common language. While some rivers may be solely referred to without the “River”, dropping it from a map, especially one which is used as a database, would not seem wise and would be an undesirable precedent given that the same logic could also be applied to any of the following: lake, lough, mountain, road, street, park, and so on.
To help us all get on the same page, I added a footnote to “Names” and a section to the Wikipedia article “Toponymy” clarifying that a composite toponym may consist of both a generic and a specific, but this doesn’t necessarily make the toponym a description. There were already some examples along these lines but hopefully it’s clearer now.
In fairness, many databases, especially address databases in some countries, try to split the generic from the specific to facilitate automated abbreviation or abbreviation expansion. However, these schemas usually have to make some assumptions that break down outside of a specific geography or cross-culturally, sort of like when programmers think a personal name must consist of a first, middle, and last name. In some languages, isolating a toponym’s specific takes on added importance for grammatical reasons. Affected language communities in OSM have adopted sorting_name=*, name:genitive=*, etc. to triangulate around this semantic boundary based on various use cases.
I’m not suggesting a mechanical find-and-replace.