Publish a new release or otherwise ensure iD can import recent updates to the NSI

[rjw62]

I see that this project hasn't published a release since v4.0.2 in April 2020. I assume that this is why iD hasn't updated its version of the NSI since that time. iD is thus not taking advantage of more recent updates to the NSI, and more importantly is continuing to entice mappers to make erroneous changes based on mistakes in that release of the NSI that have since been fixed.

I understand that there have been some changes to the output format of the NSI, which may cause problems for new releases and imports by iD. Nevertheless, I think something needs to be done between this project and iD to ensure that new releases can be made and will feed in to iD in a timely manner.

[bhousel]

I agree, we should release a new version soon. The next published version of NSI will be a major release with breaking changes, so nothing will happen really quickly.. But I do expect to be able to push an update to iD in the next month or so.

[TheAdventurer64]

Regarding this, GoMap has the newest changes to the NSI, while iD is still using outdated presets.

[quincylvania]

@TheAdventurer64 That's nice for GoMap, but keep in mind that iD's update cycle isn't tied to any other editor. In open source, things get done when they get done 🤷

[bhousel]

We're pretty close on this!
I deployed an iD with the preview NSI v5 code to: http://www.ideditor.org/nsi-v5/
Try it out and let me know if you find any odd validation or preset suggestions!

[kjonosm]

let me know if you find any odd validation or preset suggestions!

Hotel name validation seems to strict:

[kjonosm]

Validator warning generally refers to an outdated or incomplete brand (Marke in German) instead of an outdated or incomplete network or operator:

[kjonosm]

Validator still wants to update supermarket name although brand tag is already correctly set:

[kjonosm]

Nextbike operates several differently branded bike sharing systems across Europe. The validator seems to check for matching operator tags and wants to update to the wrong brand. For this example id should update to Edeka nextbike:

[kjonosm]

weird update suggestion for a stamp vending machine apparently based on matching operator, see here

[kjonosm]

Wrong update suggestion for the network values of this relation. Maybe only match features with matching network tag and ignore matching operator tags.

[bhousel]

Thanks for testing @kjonosm !

Hotel name validation seems to strict:

Good catch, I was missing the preserveNames property on these. I'll push an update to the NSI dist/* files shortly.

Validator warning generally refers to an outdated or incomplete brand (Marke in German) instead of an outdated or incomplete network or operator:

Oh yeah, this is a localization issue that will be updated on the iD side eventually.. The new English source strings are a bit broader:

noncanonical_brand:
  message: "{feature} looks like a common feature with nonstandard tags"
  message_incomplete: "{feature} looks like a common feature with incomplete tags"
  reference: "Some features, for example retail chains or post offices, are expected to have certain tags in common."

Validator still wants to update supermarket name although brand tag is already correctly set:

Yea this is a lot like the Tesco example here. Do you think people would be ok if the validator offered tagging like name=Edeka+branch=Barnimstraße? The osm wiki does encourage it, so it seems like it would be a reasonable suggestion.

(will respond to others soon.. )