Someone has changed waterway=spillway and man_made=spillway to (mostly) usage=spillway. I initially thought that it was a mechanical edit, but given the variation in the tags used and other changes, it may have had some “non-mechanical component” that introduced other errors.
waterway=spillway was documented as being used by data consumers; likewise man_made=spillway. Someone had marked it as “deprecated” in the wiki but it was still fairly widely used (whether the wiki is supposed to document tags in use or not is probably a discussion for elsewhere).
Unfortunately, it seems to have introduced a number of errors such as here and here (though other edits were correct - here). There were 300-odd changes; I’ve only looked at a few that I’m familiar with and a few near me. There may be more egregious errors elsewhere.
The difficultly with mechanical (or even just remote) tag changes to spillway keys is that it’s used for both linear and area features, and on ones where there may always be at least some water and where there is not always (intermittent), and on large and small artificial waterways (perhaps analogous to canal and drain). It has also been used (for problematic historical reasons) on spillway-like features such as fish passes, again both linear and area features. For each of these combinations, it looks like the wrong choice of target tags has been made somewhere.
Options going forward include:
Revert the changes and then reassess how best to tag example properly, in each case identifying:
Whether the thing is not a spillway at all, and if not how best to tag it
If it is a spillway, whether the combination of area and linear features at the location was correct. If not, how best to tag (say) a more often wet trickle of water over a less often wet concrete area
What intermittent tags make sense - is an “intermittent” spillway feeding a non-intermittent river? If so, how?
What waterway feature makes sense? Is canal ever correct?
Leave the changes as they are and “bucket and shovel” the data that remains so that it is “at least not wrong”, and also change data consumers that made rendering decisions based on the previous OSM tags.
Do nothing and accept the degradation of OSM data.
Either of (1) or (2) will require considerably more work than current tag changes took. In most cases it’ll need either a site survey or talking to someone who has actually been there and still has a mental picture of the area, since these are often in pretty out-of-the-way places and aerial imagery often isn’t a priority there.
My preference is option 1. In any case, the usage=* key is overused, and it makes little sense to me to mass edit more tags into it. If anything, I would like to see objects from there retagged to other keys.
Since we’re talking about spillways, you can start by verifying the ones near dams as “probably correct”.
I don’t understand what that paragraph is trying to say? Simplifying hugely, there are basically two sorts of spillways associated with hydro power schemes:
Usually small spillways downstream of e.g. an Archimedes’ screw
Often large spillways that prevent water from overtopping a dam.
Is that paragraph talking about only one or both, and what about non-hydro power spillways? Perhaps you could rewrite it using a few more words to explain a bit better?
The proposal’s paragraph dealt with water courses intended to channel overflowing water from any place (hydropower or not).
Do you refer to this kind of thing?
To me on this picture, the intermittent spillway is on the left and the Archimedes’ screw on the right is whether a generator or a pump (more likely a generator here) and they are two different waterways with their own OSM feature.
By the way and regarding the main point of this topic, I’m not encouraging automated edits and every feature should be reviewed carefully before replacing tags. Unless it has been discussed before and agreed, mapper should now handle error cases with care (or unfortunately be reverted).
Yes - exactly that (for info, here in OSM). The thing at the left is a regular lake spillway and often dry; the thing at the right is below an Arhimedes’ Screw, not currently mapped in OSM, and usually wet.