In an ideal, sane world, it wouldn't be one. Yet, draconian copyright is a minefield which must be navigated / dealt with / at least avoided.

While, to my non-lawyer understanding of copyright, facts (themselves) don't qualify, that doesn't stop the source from disagreeing and bringing legal action anyway. It's a risk. One most easily dealt with (by OSM) by deleting any suspect data (which would likely include changesets that don't cite a source).

Besides, these days collections of facts are subject to copyright (encyclopedias, maps, to name only 2).

“A quick Google search will reveal that both 15 and 17 are still used by the Post Office”
Same questions apply; what's the source, and what's the license of that data?

If actually Google, then definitely risky.

Copyright being inconvenient doesn't mean it can or should be ignored.

Compare, for example, OpenFoodFacts; there is an important difference between taking a photo of product packaging (for the info printed on it) versus copying the same data from other sources (like the manufacturer's website, or a database). Photographing packaging is rather less legally risky, whereas copying from elsewhere is dubious at best. Arguments of ‘but it's the same result’ are irrelevant (and legally untrue; not the same result, since one involved copying from a proprietary source, and the other did not).

In OSM that'd be akin to the difference between surveying versus copying some other source. A survey is the surveyor's own work, unlike copying from another source.

Shortcuts, while they may sound rationally reasonable (because same result), aren't so simple legally. In copyright there is much weight put on independent authorship versus copying. Regardless of how similar the result is; that's not the point, but the method of having something similar is very much the point.

“Almanacs of street addresses in the library, annual publications going back to the 1800s”

If those were the sources (and they're that old) then they're likely now in the public domain.
Yet, you indicate that they're not the sources, a much more recent publication is (for which, by default, copyright re-applies). So, this whole line of reasoning doesn't hold up legally.

What matters is from where the copying WAS done, not where it could've been done from.

The method is different, and that matters.

Just as taking a shortcut through my neighbours back yard may get me home quicker, and has the same result (for me) as using only public roads; doesn't matter, I would still be trespassing on their property instead of taking a legally sound route.

“Can't see it's an issue.”

Doesn't mean that it isn't one. Being blasé about it doesn't make the problem go away.

Besides, the matter seems best decided by the OSMF or a relevant OSM WG. Better to check & be sure before legal trouble, than to receive a Letter Before Action.

I'm not making excuses for draconian copyright, and am not saying that this is how it should be. But, that's how it is, regardless of what might be sane.

Are you sure the tagging is correct? There are some ways with name=Palace Close which aren't joined to other ways with the same name, but are sandwiched between ways which you did rename.

source?

source?

What makes for osm.wiki/Good_changeset_comments 🙂.

I had deliberately left the original PoI node in place:
▪︎in order to osm.wiki/Keep_the_history
▪︎because while the ground floor (level=0) is a supermarket, level=-1 is parking and level=1 is houseware retail (though, it's all Waitrose)
▪︎I wasn't sure (especially at the time) of the best (or least-bad) way to map the true arrangement; maybe the non-building: tags should be removed from the building outline and there should be several PoI nodes (I'm intending to (re)survey the site when I get chance). Thus, for the time being thought it best to preserve elements rather than hastily delete (to then decide from surveying).

If I recall, the tagging was non-identical, deliberately.

Accuracy over neatness.

This whole site needs an overhaul; lots of missing features. Hence intending a through survey (but, that won't be quick).

👍 thanks (from a Channel Islander). I saw the same video, and wondered if OSM was up-to-date.

I didn't know that the Port of Dover had expanded to cross the Atlantic.

FYI: composing a response, which became & took longer than expected. Hope to submit it either later today, or tomorrow.

After doing some digging (after dealing with some other distractions), this may be a problem, and may need the attention of probably the License Working Group.

The closest thing I could find was http://www.theislandwiki.org/index.php/Copyright_notice but that's not a license, more of a disclaimer. The statement that the authors believe something to be public domain doesn't make it so.

Even if so, then copyright may still apply to The Island Wiki's collection of data, and need a suitable license (or explicit permission) to use within OSM.

While their claim that facts aren't subject to copyright is, as far as I (a non-lawyer) know, correct. Claiming it doesn't make it so, especially not when there are actual legal stakes and the answer really does matter.

While I still think it would be good to bring this to the attention of relevant Working Groups, if only for due prudence, the simplest path might be to get express written permission from TIW, which is then posted in a Jersey-specific section of OSM's wiki.

Keep in mind that they've not said that their own work is in the public domain. By default, without having to do anything (see the Berne Convention) a work is auto-copyrighted the moment it is authored. Thus, by default, it must be treated as being proprietary.

Sigh. I dread to think how much has already been copied. Even identifying all the changesets (since many lack source=*) will be a challenge. The risk is, now, that much of the data for Jersey may have to be reverted. Fun.

Only just become aware of your reply.

Ah, The Island Wiki. Knew of the site, for different reasons. Thanks for the pointer, regardless.

However, what's the copyright license of the data? We (OSM) can't copy from sources that have proprietary (or otherwise incompatible) licenses.

The northern cliffpaths are overdue some much needed attention.

Would've been good to set source=GPS based on the comment=* .

Better still would be to upload the captured trace; we're lacking in those, too.

At some point I think we need a conversation about how to handle Jersey's sub-divisions, given the differences with larger jurisdictions.

There doesn't seem to be any obviously good scheme that I'm aware of.

But, more obvious features, first.

Who knows, maybe we'll get it done within our lifetimes 😋.

Holy crap; you've had a busy & productive evening 😀. 👍

Pity that I'm now a bit too far to easily survey these.

“I just copied the original”

I suspected so.

I left a comment on the original changeset (which introduced the suspect key) with some advice for traffic_calming=* (since there's a LOT of it here; local government seems to be anti-car), so will hopefully be avoided in future. There's enough messy data on this island already.

“I just changed it.”

Saw that changeset. 👍

👍🙂

Some feedback, since you're relatively new.

source= ? Particularly for purposes of verification.

For entrances, they're assumed to be 2-way by default, so no need to add (entrance|exit)=yes . The exception is when some are (entrance|exit)-only (some larger establishments have oneway door systems).

For exit-only (e.g. node #9064598814 which has entrance=no & exit=yes) then simply entrance=exit . The typical use of exit=* is exit=emergency . See the wiki.

The entrance=* tag should denote the purpose / significance of the doorway. E.g. entrance=main (vs. entrance=service).

For access=permit , according to the wiki this is for when a permit is fairly easy to obtain. When permits are difficult to obtain, then the recommendation is to set access=private . Though, that may match what you intended; just pointing it out.

air_conditioning is a key I'll keep in mind for future 🙂.

As for the way (#179245903):
▪︎opening_hours:covid19=same — same as what? There's no opening_hours=*
▪︎operator=Government of Jersey's Modernisation and Digital Department — to avoid having a thousand values for the same entity (think of data-parsers), simply Government of Jersey . It would be good to also set operator:type=government in such cases. Context of which section of the States can be derived from context (such as where website=* points to, plus the value you set for description=*). Else, make description more specific: description:operator=Government of Jersey's [… whatever …].
▪︎website=https://www.gov.je — firstly the string should have a trailing slash (to be a properly formed URL). Since this seems like a major office for this section of the States then a more specific URL pointing to the website section for said department. Preferably a short one which will continue working in the future.
▪︎addr:city=St Helier — should match (in this case) the short_name of the containing administrative boundary, which would be addr:city=St. Helier . It's an abbreviation, after all. Else use the expanded version: Saint Helier . Arguably the name=* of the admin boundary should match the short_name since that's commonly used, despite being an abbreviation (which are discouraged).

I think that was all. Thanks for your contribution.

You might also consider kerb=flush if that's also the case for a given table.

Re node #9038918986 (the crossing (outside the school) which is also a traffic (‘calming’) table): see changeset #110667321 .

In future, you want to set traffic_calming=table for such things (or possibly traffic_calming=hump, depending on its shape).

See traffic_calming=* .

traffic_calming=raised_table is only used 21 times globally (unlike the recommended traffic_calming=table which occurs 120,077 times).