For buildings in Switzerland, tags with EGID are helpful. Currently, we use “ref:gwrid” and “ref:EGID” for this. A standardized format would be “ref:CH:egid” (the tag being mostly used in CH). Key:ref:EGID - OpenStreetMap Wiki has some description.
Currently the use of both tags is fairly limited (ca. 1500), but potentially all buildings in CH could have that.
We don’t in the generality you are implying, 1’500 is nothing.
There is no need to prefix references only used in Switzerland with CH as they are only used in CH (yes it is a common anti-pattern in OSM). Uppercase in keys is a further anti-pattern.
But in general we should be keeping external keys on OSM objects to a minimum, 20+ years of experience shows that they end up all over the place without explicit editor support so the question should be more: is there enough (actual) value-added in adding the egid to buildings (noting that the definition of what a building is might differ from the GWR and that might require changing how we map things) to outweigh that pain.
“we” are the mappers who use the identifier with buildings in Switzerland. It’s not meant to include you or to refer to you.
There is a regularly run report that largely relies on addresses for EGID-identified buildings in Switzerland. Unfortunately the report doesn’t include the identifier nor is the report community maintained. I’m not sure if it’s even open source software.
I am convinced it is a bad idea to start systematically adding EGID references as keys to the OSM database, in general echoing the reasons mentioned by @SimonPoole in this thread.
Another point:
The answer to the question “What is a building?” as answered by EGID/GWR is only one answer to that question, with their answer largely being informed by an administrative perspective, of course. Other people will have other perspectives of what constitutes a building and thus, linking OSM buildings to EGID entries will enforce one of the possibilities of how buildings are mapped in OSM. Two examples:
The Hotel Les Trois Rois in Le Locle is represented in the GWR database by five different EGID values: 1475529, 1475530, 502354269, 1475531, 502354489. But if you look at the building (or go to it), would you consider that being multiple buildings: File:TroisRois 20260118 162339.jpg - Wikimedia Commons ? There is a clear architectural idea behind this (which translates into actual usage), and therefore it is mapped as one structure in OSM. This is not to say that perspectives that consider this multiple buildings aren’t valid, it’s just to say that there are multiple ways of looking at the built environment and EGID/GWR is only one of them. Which EGID would you map to this building? All of them? Split up the polygon?
Another building would be EGID 2084823, which is represented in OSM as one single polygon (because it was imported from cantonal open data): Way: 147960688 | OpenStreetMap . But I could also argue that representing this way as different polygons would actually increase mapping accuracy, given the very different structures that this way currently represents: One end of the building is here, the other one – including considerable changes in topography – is here (the one with “Walder” written on it). If somebody now wants to add more metadata to one of these buildings, e.g. architectural style, and would split the polygon, would you keep the same EGID twice? Not incorporate other metadata that cannot really be added while keeping true to the EGID database?
This is not to start a debate about how these two cases should be handled, but to illustrate how EGID values might correspond not as directly to OSM as you maybe expect.
Sorry if I went off topic (choice or naming of key). Feel free to message me on www.osm.org or open a separate topic if you want to discuss related, but different topics.
The main problem with using “gwr” in the key name (or as key) is that WR/RegBL/REA (or housing-stats.ch) includes other identifiers, even if EGID is the main identifier.
In theory, one could have suggested simply “egid” as key for whatever counterintuitive reason. However I got the impression that modern keys don’t follow that naming.
Maybe ref:CH:g or ref:CH:b could be alternatives (if E stands for CH and ID for REF), but that seem prone to confusion.
Most of the ref:gwrid tags were added by me. I used this one because that’s what the field was called in the source. I had used it to track changes in the source.
I stopped doing that after spalinger (e.g., in Changeset 159516043) deleted most of the addresses and entered them as a separate node (sometimes outside the house). The gwrid stayed by the house, which doesn’t really make sense. So, I suppose these should just be deleted.
Not a fan of the ref:CH:egid, same as @SimonPoole .
IMHO you were tagging the EGAID … the address id, not the building id, I could get more behind doing that from a usefulness pov, though I’m not really convinced.
The standard would have been ref:CH:egid. Luckily nobody wants the status quo. It took me a while to even notice that we had a second key for the same.
Is “never miss an opportunity to pontificate” the forum motto?
As Simon mentioned above, prefixing the egid reference key with CH is not necessary.
I’d keep it simple, and just go for ref:egid.
Is “never miss an opportunity to pontificate” the forum motto?
No, as this would be unnecessarily negatively connotated.
If you think I’m expressing my opinions in a self-important way, then you can simply say so, and I’ll take care to be more concise.
I think people including yourself had ample time to comment and participate in the pool. It would be bit weird to just repeat what you already stated/polled before.
.JPG){kind=link}
_(likesopenstreetmap)_2017-05-12_15H43M39S000.jpg){kind=link}