Map Maintenance: Handling of fixme=name and suspicious combinations of name=*+noname=yes

[kmpoppe]

Situation

There are currently ~160k objects tagged with fixme=name.

Understandably, SC currently ignores fixme tags completely - IMHO most values of that key would be impossible to survey without deeper investigation.

Yet, fixme=name seems like a very straight-forward request: "I (the last mapper) was unable to determine what the name of this feature is, please help out".

Proposal

I therefore would suggest for SC to work the following way:

a) remove fixme=name (if present) when the user answers quests/road_name or quests/place_name with either a real name or noname=yes (because otherwise, things like this (noname=yes set, fixme=name persited) or this (name set, fixme=name persisted) happen)

b) the aforementioned quests should force a survey of the name when a fixme=name is older than one year to get rid of it, because all viable answer paths from there (name=*, noname=yes or changing the type of road that doesn't get asked for a name again) lead to a situation where the fixme is no longer relevant.

c) FIXME=name (1.5k objects) could be handled the same way.

SCEE

I do not think that this option should only be something relevant for SCEE, because the handling is purely under to hood and there's no expert knowledge necessary to answer those already existing quests.

Feedback appreciated !

[matkoniecz]

I am tempted to propose bot edit that would remove fixme=name from objects without any name tags but I admit that right now I have not time to do this.

[matkoniecz]

from https://taginfo.openstreetmap.org/tags/fixme=name#chronology it looks like it was brought in large parts by automated mass edits and there is potentially organic growth that stopped and got reversed

so hopefully once cleaned they would not return in meaningful numbers

[kmpoppe]

I have now checked the history of all 15,459 objects, that have (fixme|FIXME=name|name?|Name|Name? and either have name=* or noname=yes) whereof 4,705 (30.44%) had the fixme added before name or noname was set - those might be objects were I will ask the forum and tagging-ml if somebody would oppose an automated edit.

For the other 7 out of 10 objects a manual check would be needed, as the fixme was added, after name or noname was added.

Nevertheless, 145k objects remain with a set fixme without a name or noname. So, if ever those were edited with SC, the fixme must be removed from hereonout (point a) of my list above) - do you agree?

[kmpoppe]

In the same way, sollowing @mnalis's suggestion, 13,055 objects have name=*+noname=yes, 9,923 (76%) had their name set after noname was set and they should therefore be included in an automated edit to remove noname.

Thoughts?

[mnalis]

@kmpoppe both automated edits as well as SC change to remove fixme=name (and its ilk) when name is set sound good ideas to me!

[kmpoppe]

Seeing the replies in the community, the list(s) and the wiki discussion I feel like automation is not the way to move forward, but rather going the MapRoulette way where every object is manually checked before changing.

Am I the only one feeling that?

[mnalis]

Makes sense to me. a) (and by extension c)) especially sounds like good stewardship from SC side, as there are about 160k of fixme=name tags!

b) also makes sense to me, to proactively resolve those ambiguous situations. I though to suggest that might be helpful if quest pre-filled existing name if present, but that sounds like it would create more confusion than it would help, and also might lead to people incorrectly confirming name which is not correct... So probably better to ask name again from scratch..

In addition, name resurvey could also be applied to name=* + noname=yes combinations (same 13k matches) and get rid of one of those tags...