Particularly without peer review / community discussion. For example I noticed the tag education=school, approved through voting earlier this year, that was already in very modest stagnating use for over a decade has recently demonstrated the power of exponential growth, and I grow the suspicion iD might be involved in this unexpected explosion:
I have seen newbies with the experience of just a few edits uploading changesets where all they did was adding the tag education=school to a school already tagged with amenity=school (and an id changeset tag about updating “outdated or incomplete tags”).
From what I understand, the community in general wants this change. And this is the best way to achieve it because it gives tag users and data users time to adjust to the change
Possibly because Editors by nature allows user to perform Automated or semi-automated edit, and their development would possibly stale if following Automated Edits code of conductfor each and every change in presets or validation rules.
it can have similar impact in the long run, but presets just come to effect when you search for them, these “ improvements” are active and asking to be applied as soon as you have eligible data in your editor (correct me if I’m wrong, not using iD a lot).
I don’t mind it. I’ve added it to 2,000+ objects (I was doing something else at the time) and it has shown me how awful the mapping of some schools is locally, e.g. amenity=school added as a node, the building AND the site, all for the one school.
If I map education=school, shouldn’t iD prompt / automatically add amenity=school? Like what happens with healthcare.
Also, on the standard view, landuse=education should show as a similar colour to amenity=school. landuse=education is useful when one is dealing with multiple schools on a single site.
I didn’t make a thorough analysis, maybe there’s a different mass edit as well, maybe it really is just natural growth because the tag passed the wiki vote, but to me the cases I looked at seemed like automated id edits (“suggestions”, which a newbie user hardly can understand)
It was iD that was prompting me, but I did try to at least add school=* which made me stop for 5 seconds to verify it is a school, it isn’t disused, it isn’t a college and that the basics were right, e.g. it wasn’t amenity=school name=Tesco.
“what happens with healthcare” illustrates the problem nicely. Sometimes only healthcare is set, sometimes only amenity. Sometimes both keys are set but to different values.
It is possible to try and make sense of it, but not easy.
I think it is not ok that iD suggests such tag upgrades in general. iD is used by newbies who cannot judge whether these suggestions make sense in general or in individual cases. Such suggestions can be made in a QA tool used by people who know exactly what they are doing. If it is absolutely clear that a change is always useful (e.g. in the case of spelling mistakes), then it makes more sense to do it with a mass edit, because then no one in iD will be bothered by it unnecessarily. I think this function can easily be used to push through controversial changes if only the iD maintainers are on your side. This is also the case with presets, but not to such an extreme degree.
If iD still wants to offer such tag upgrades, then this should only happen in an expert mode that you have to activate first if you consider yourself qualified to do so.
This was already covered. But I think it’s the key misunderstanding here. There are some warnings that can show up as soon as they’re in your view, but only when you’re about to submit your change set. Mostly you need to be trying to edit the object to see these warnings.
Regardless, this change is only doing exactly what the proposal allowed. A proposal is a far more detailed change log than an automated edit document. It’s not a destructive change either. Can anyone point to a single incorrect change caused by this editor warning?
Additionally, a user had to engage to make the change. Whatever you think of the change, this type of thing isn’t automated - somebody had to click on the object, see the warning, then apply the change. I can do more damage incorrectly applying a preset than with this warning.
Lastly, the iD maintainer is an OSMF employee, so worries about only getting them on your side aren’t especially compelling to me. Especially when we’re talking about an approved proposal. The proposal got 75% of voting community members in it’s side. Far from only the iD maintainer. If the iD maintainer did go rogue - and they haven’t - OSMF could hold them accountable. They did their job here, helpfully and correctly.
So, to sum it all up, I think iD did the right thing here. If you disagree with the change, I think you should make a new proposal.
All a proposal means is that (in this case) 27 people who read the wiki thought it was a good idea.
If it introduces data into OSM that contradicts other data like this list and this list then it is a very real problem.
Have you seen the garbage changes added by people “agreeing with iD’s recommendations”? Do you really want a list of those?
It’s perfectly sensible for iD to say that it is detected a problem (and busywork such as adding useless tags such as this one is a bit of an irrelevance compared with the main problems caused by iD), but in most cases the way that iD suggests that tags should be “upgraded” is a falsification of the actual situation.
27 people is a lot more than most automated edit documents get. Are you suggesting the proposal process shouldn’t mean anything? I’m not trying to be hyperbolic here. If iD can’t offer a minor tag adjustment to a user after someone does the work of accepting community feedback and iterating on a proposal, then why even go through the proposal process at all?
And at least now we’re having a conversation about ways to possibly improve the editor warning instead of calling it things it isn’t.
With my question you responded to here, I’m not referring to editor warnings in general. Just this specific healthcare warning that people are concerned about in this thread. Can anyone point to one that was incorrectly applied?
landuse=education is useful when one is dealing with multiple schools on a single site.
