Proposal: automatic retagging of waterway=soakhole

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa · 2024-02-17T07:32:28.688Z

Hi, 14 years ago the tag waterway=soakhole was invented and used in an import in New Zealand. Recently, a new tag was invented with the same defintion: waterway=stream_end. This new tag is now more popular (see graph), so I’m proposing to do a mass-edit in New Zealand to replace 4200 nodes tagged as waterway=soakhole with waterway=stream_end.

In addition to the 4200 nodes in New Zealand, there are also 53 in Australia and 10 in the rest of the world. At this point, I am not proposing to update features in any other country besides from New Zealand.

I am not aware of any data consumers using the tag waterway=soakhole, and it is not supported by any editors.

Would this kind of mass-edit be acceptable? Any thoughts or advice would be appreciated. Thanks

TheSwavu · 2024-02-17T09:17:46.726Z

I would prefer that the ones not connected to a waterway are deleted, rather than just modified.

There are currently 4267 waterway=soakhole in NZ but only 179 of them are actually connected to a waterway overpass turbo.

It was an odd decision to import the ends of the waterways before the waterways.

Kovoschiz · 2024-02-17T10:40:00.462Z

Why can’t they indicate indicate missing waterway= that needs to be drawn? There’s no requirement to draw the =waterway or =pipeline for a =inlet / =outlet either.
I prefer =soakhole , as it looks clearer than =stream_end to me. It also resembles natural=sinkhole for the style. If they can exist for =ditch, it avoids the need for =ditch_end , or counterintuitive use.
However in rainwater drainage, there are soak pits / soakaway. The two needs to be considered. They may be used with septic tanks as well.

TheSwavu · 2024-02-17T21:46:36.535Z

waterway=stream_end is the equivalent of noexit=yes, that is it indicates the end of a waterway so that QC checks will know that the end of the waterway is intentional and not a mapping error. =inlet / =outlet are physical things and if there is a soak hole the correct tagging should be something like natural=soakhole. Although it would need a better name than “soakhole”, because this name is also used in NZ for constructed stormwater management infrastructure Everything You Need to Know About Soakholes - Drainage NZ.

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa · 2024-02-19T10:36:59.309Z

TheSwavu:

I would prefer that the ones not connected to a waterway are deleted, rather than just modified.

I think the reason no one has deleted them in 14 years is because one day we might finish the waterway import, and then every node should be connected…

I didn’t want to propose something as radical as deleting all these nodes. I was just hoping to avoid a situation where there are 2 popular tags with exactly the same meaning.

Kovoschiz:

I prefer =soakhole , as it looks clearer than =stream_end to me.

It seems like the rest of the world is strongly in favour of =stream_end, despite =soakhole already existing for a decade before =stream_end was invented. As @TheSwavu mentioned, the term ‘soakhole’ can also be confusing since it has 2 meanings.

I personally think it’s better to change the weird New Zealand tag to the more popular international tag. I’ve actually started a list of all the bizarre tags that the original import invented.

TheSwavu:

It was an odd decision to import the ends of the waterways before the waterways.

I wasn’t involved in the 2010 import, so I have no idea why it was done like this… Maybe to satisfy a validator?

Anyway, it would be nice to hear from anyone else who has an opinion. Maybe the first step would be to mark one of the tags as deprecated on the wiki (?) so that it’s obvious to mappers which tag is preferred

Kovoschiz · 2024-02-19T12:15:07.892Z

They are 2 aspects. How about =stream_end + natural=soakhole (to be improved, but looks acceptable with natural= ?) mentioned above. This allows =stream_end + =sinkhole + sinkhole=ponor for the other case.
There are currently 157 =sinkhole on ~220 =stream

TheSwavu · 2024-02-19T21:19:25.641Z

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa:

I didn’t want to propose something as radical as deleting all these nodes. I was just hoping to avoid a situation where there are 2 popular tags with exactly the same meaning.

The reason why there are two tags is because people did not want to use soakhole as it is badly defined and only came from a single import see Support `waterway=stream_end` by Famlam · Pull Request #2034 · osm-fr/osmose-backend · GitHub for example.

This was a poor quality import and tag-fiddling is not going to fix the problem.

IanH · 2024-02-20T01:14:04.593Z

I think the decision between =soakhole and =sinkhole should be based on depth. The average adult should able stand in a soakhole without being fully submerged. This would make anything deeper than 3m be considered a sinkhole. Though most sinkhole are often much larger than that.

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa · 2024-03-19T08:25:06.935Z

thanks for the advise, for now, I’ll only update the wiki page for waterway=soakhole to suggest that people use waterway=stream_end instead. This way the 2 tags won’t continue to grow in parallel. I’ll won’t any update any of the half-imported features in New Zealand.