I think iD is really ok nowadays.
In the beginning it was not very good. But now you get good suggestions for tags and there are only a few geometric tools missing. The only thing making me use Potlatch 2 sometimes is that you can use GPX-waypoints there. I haven’t found that in iD.
I use JOSM only for revert or imports (which is something I don’t do often).
Ok, I’ll give it another go. I mostly use JOSM with filters and overpass queries but for simple edits like adding a missing lake it seems quite quick and easy.
iD pales in comparison though. I think especially in creating good geometry. Because iD becomes laggy and difficult when there are too many nodes - and it doesnt have a refining tool it just becomes more convenient to leave the geometry at a very rough level.
Today I created this for somebody in Italy. Such geometry would have taken a lot of time in iD, but only about 5 minutes in JOSM.
I am happy about every user that starts to use ID. They understand att least that they are ediding att database.
A much bigger problem are even more simplyied apps. Streetcomplet for example adds curbs on footwats marked som “informal” and asked for övergångsställen where a hiking path is corssing a rural road. But users do not understand what the question meens in relation to other objects.
But they do improve Openstreetmap even if much is not perfect. And it makes clear for oss that we have a need of simple applications.
So the main focus should be to support new people. And their is a riskt off Bering to complex if you suggest JOSM without good support.