As far as I can see this seems reasonable and making sense and representing a real feature, which would be useful to have.
Code also looks reasonable.
I am asking for review/feedback/protests as I am not really familiar with such feature (not seen one in my life, I seen some disappearing streams but none at a stable position). It also has no sky-high use and has not passed proposal process. But neither is some fundamental problem, as it has some decent use.
I am currently planning on merging this pull request which would add it to iD presets.
Is there are any reason to not do this? Are there any serious problems with this tag?
there was some time since my previous attempted feedback-gathering thread and this is much more focused, so I hope this one will go better than the previous one
if anyone would be interested in promoting this one as alternative (RFC? proposal process? retagging as discussed with local communities and preferred by them? Contacting users of that tag) I would be happy to wait to see how community would react to the attempt (please, comment in this thread if you plan to do so)
but as it stands I cannot justify pushing it over something used by mappers
I have not come across this tag yet and checked a handful of those used in Germany. None of them fits the description
The natural end of a stream (where it seeps into the ground).
A dozen or so are used to mark the points where ditches along roads are branched off to discharge down a slope or into a grass patch. Others are used to mark the point where a stream enters a pipe and continues underground.
I think this is due to the poor definition “stream end”. From that point of view I agree to @phodgkin that waterway=soakhole describes the end of a stream where it seeps into the ground much better.
Besides that according to my own experience stream ends are in most cases not point-like objects. Very often the water seeps into the ground in a swampy area or moves back and forth in the streambed depending on the water level of the stream. Cases where a stream soaks into the ground continuously in one small spot may be common in certain areas (depending on the geological conditions), but in many places are not.
Having said that I think stream_end end soakhole could well describe different situations.
I think before adding stream end to the iD presets the wiki page should be updated to give proper information about the different kinds of “stream ends” and also about those cases when it should not be used.
If that would be anywhere close to general use of this tag then its claimed description completely mismatches actual use. In such case it should not get into presets in its current state.
And admittedly name is asking for degeneration into “waterway here ends in some way”.
I don’t know this place but I strongly believe this is one of those very tiny springs where a small quantity of water trickles into an artificial samll basin for hikers to refresh. There are hundreds of those in the mid German hills and usually the water is routed through a tiny ditch or pipe, ending at the next downhill slope where the water just vanishes. I would not call this a stream and very often these tiny springs have dried completely within the last 10 years.
I have not taken the time to start writing notes to those “stream_end” I have roughly checked because I’d say it’s a poor tag anyhow. I understand it exactly as
which should be reflected in the wiki page, whereas waterway=soakhole clearly defines a spot, where a stream (not a ditch or artificial waterway) naturally ends by definitely seeping into the ground.
I’m not so keen on this value, since years as already been spent to clean and reduce the amount of values for waterway=*.
We also have fairly more established and better defined tagging like natural=sinkhole and inlet=* that would deserve to be used.
We may miss some “stream end” situations, like ground permeation that could be covered by a formal proposal (and probably won’t lead to a new waterway value).
Talk page of this value also shows that consensus is not yet reached and it would be suitable for a formal discussion. We are far beyond the point of discussing a preset, prior reaching a consensus on its actual extent.
What guidance will be given to iD users to help them decide whether it is an appropriate tag or not?
The wiki page is a bit vague - is it for geological features that don’t move (e.g. limestone like Water Sinks near Malham in Yorkshire in England) or vague swampy or other porous areas? If it is definitely not for “streams mapped from aerial imagery that just appear to end”, how will iD say that?
I’ve done a lot of stream mapping in Belize, and wherever the water disappears into the ground, I use waterway=soakhole if it’s swampy and waterway=sinkhole when I’m fairly certain it flows into a cave like hole.
In the uk a stream disappearing into the ground in karst scenery is called a “sink” (probably similar to your “Soakhole”). The location where it first emerges is a “spring” (or annoted with “Issues” on our Ordnance Survey mapping. We also have intermittent streams which depend on the ground water level at the time so seasonal variation should be considered. In some cases (ie in chalk) water can exist in some sections but not others but in times of high rainfall the stream becomes continuous. Perhaps it should be applied at a point beyond which water has never been observed to flow.
I prefer mention of sinkhole or soakhole/away. One thing such a point is not is the end of a stream. Just because you can’t see the flow, does not mean it does not exist. Drying up in low flow is something else.
It’s not “ways better”, it’s actively misleading. According to the Wiki (as adopted from Wikipedia):
A sinkhole, […], is a natural depression or hole in the surface topography caused by the removal of soil or bedrock, often both, by water and/or by collapse of an underlying cavity. A typical form of a sinkhole is a crater.
In general (and as many counterexamples, many common to UK, would confirm), a natural=sinkhole does not even typically occur in relation to a waterway. Indeed, an overpass query in karst areas of Dalmatia, Herzegovina and Montenegro shows many that are just craters, not connected to any waterway, like the one from our article.
In contrast, the one I surveyed and mapped (as a waterway=soakhole, the only tag I could on Wiki at the time), located on the Sušica river in the Durmitor National Park in Montenegro, looks like a rather dull pile of rock:
As pointed out above, that’s usually not just a single “hole”, but a long stretch of porous riverbed that absorbs the water to underground channels and/or caves. Still, it’s useful to explicitly map a certain point where the bed becomes permanently dry (just as we map noexit=yes to affirm the road does not continue here).
So yes, those things should be under the waterway= space. I’d prefer that we adopt the soakhole since that’s all we currently have (and stream_end is vague beyond repair).
