Changeset: 108430155
updated color=* to colour=* and fixed all errors and some of the warnings I found
Closed by Friendly_Ghost
Tags
created_by | JOSM/1.5 (18004 en) |
---|---|
source | Mapbox Satellite |
Discussion
-
Comment from Lee Carré
Holy smokes, Casper; smaller changeset areas, please. Otherwise you're liable to scare the fleshies.
-
Comment from MoiraPrime
This changeset area is 2 big. I'm scared. 😱
-
Comment from Lee Carré
See, what did I tell you, Cas?(!) 😄
-
Comment from Lee Carré
PM'd Friendly_Ghost with a pointer to here, since he's busy editing and likely hasn't seen this discussion, yet.
-
Comment from Friendly_Ghost
I only just noticed, lol. These bboxes need to be big so I can save myself some time.
-
Comment from tekim
Please!, find a way to group these changes by geographic area, and keep the bboxes a reasonable size.
---
Published using OSMCha: https://osmcha.org/changesets/108430155
-
Comment from Friendly_Ghost
That would take me 4 to 6 times as much time with no real benefit. I would understand this request for smaller changesets, but the ones I made yesterday were just too big for that.
---
Published using OSMCha: https://osmcha.org/changesets/108430155
-
Comment from Lee Carré
‘What's best for me’ doesn't seem overly compatible with a collaborative community project.
The obvious counter-argument is that while your time might be saved, this is at the cost of other contributors (many who didn't comment still had to mentally filter-out this changeset).
Quality takes care.
Find a way to automate your desired goal; have the machines do the sub-division work for you. Best of both worlds, then.
-
Comment from Friendly_Ghost
I can't code, so automating this process is out of the question for me, but I have figured out a workflow with taginfo, overpass turbo and JOSM to update deprecated tags reliably on a large scale.
If you have the skills to automate this process with smaller bboxes, I invite you to create a bot like that, so we can feed it deprecated tags to update.
-
Comment from tekim
Can you share your workflow with us?
-
Comment from Friendly_Ghost
Hi Tekim, yes I can. I will do that when I'm at my desk.
-
Comment from Friendly_Ghost
Workflow is as follows:
1. Browse Taginfo & Wiki for common tagging mistakes and deprecated tags.
2. Run a global query with overpass-turbo.eu, in this case nwr[“color”]. (If tags have too much or too little usage for a global query, use a bbox or country search area instead.)
3. Load in JOSM.
4. ctrl+F --> color=* and colour=* --> remove color=*.
5. ctrl+F --> color=* + “building:colour”=* --> remove color=*.
6. ctrl+F --> color=* + building=* --> update color=* to building:colour=*
7. ctrl+F --> color=* + “building:part”=* --> update color=* to building:colour=*
8. ctrl+F --> color=* --> update to colour=*
9. Press upload just to run the validator.
10. Update some outdated roof tags (like building:roof:shape=*) in the same way with ctrl+F.
11. Fix other issues that the validator picks up.
12. Run the validator again to check if the fixes work.
13. Upload all changes.All in all it takes a couple of hours to do.
-
Comment from G1asshouse
I find your mass change of one tag to another troublesome. Since both spellings (color/colour) are correct, I assume your changes are based only on forcing uniformity. As much as that is a useful goal, the OSM community has for a long time expressed that it values flexibility and mutability of the tags over the perfect conformity of the tags.
Without communicating with the OSM community at large, it is not possible to know if the “color” tags were tagging mistakes or a user’s preferred tagging scheme. If they were a preferred tagging scheme, then a conversation among the OSM community needs to occur to decide if we live with both spellings or if, as a group, we override the few editors, who prefer “color”, and force the change to “colour” only. Your actions make it appear that you have misinterpreted “depreciated” and “discouraged”, which is used as heavy social pressure and not to mean “these are mistakes that no one has yet taken the time to fix”. -
Comment from snoozingnewt
So there's actually been a bit of discussion on the OpenStreetMap World Discord Server and the general consensus is that the correct spelling is "colour". It's not like the edits went ahead with no discussion and no approval.
-
Comment from G1asshouse
I respect you and anyone that enjoys the community side channels. With all due respect, those are not official OSM channels and the final discussion needed to occur on the mailing list and OSM wiki. It gets very frustrating when folks conflate the many OSM community side channel with appropriate, community agreed upon OSM discussion methods/sites.
For those interested, here is the first mailing list conversation, I have found, related to this user's changesets.
https://lists.openstreetmap.org/pipermail/talk/2021-July/086817.html -
Comment from Friendly_Ghost
Optimistically speaking, even if a few hundred people are subscribed to the mailing lists, they still don't represent the entire OSM community. From that POV, the Discord server does a better job, even though it's a side channel.
The Wiki already lists color=* as deprecated and searching for key:color redirects you to key:colour.
We are making geospatial data so it can be used. Using data is easier when there are No multiple tags that mean the same thing, especially if one is barely in use compared to the other and is also marked as deprecated. I set out to fix that and make it consistent for everyone who wants to use colour data. You're very welcome.
The automated edits code of conduct lists obvious typo fixes as acceptable usage. That's what I did.
I'm not presenting a full plan to the mailing lists either when I fix highway=residental or =Residential. That level of bureaucracy would stagnate OSM completely. Changes like this just need to be done.
-
Comment from G1asshouse
The mailing list is the agreed upon final community discussion location. Just because you've isolated yourself with like-minded people, on discord, does not make it an appropriate alternative to the mailing list. You are welcome to bring this point up to the mailing list and state your logic for using discord instead. As of now, contrary to your wishes, Discord and other side channels are inappropriate.
My point that you do not know if there are OSM editors whom prefer to tag with color instead. If it is not a "typo" then it's a valid tag...even if heavily discouraged. I wish to emphasise that you have not done due diligence with the larger community and may have just blindly robed another OSM editor of their right to make reasonable tagging choices, especially if those choices differ from the community as a whole.
This whole thing smells a lot like "it's better to ask for forgiveness than approval". I don't think you've made these large changes out of ignorance of the community standards. I get a strong feeling that you are very aware of your actions.
If I am correct, then I find your destain for the larger OSM community to be reprehensible. -
Comment from maraf24
That's funny: https://www.openstreetmap.org/way/616060622
What's the point in correcting the name of "color" tag for leisure=park? -
Comment from Friendly_Ghost
@maraf24 ¯\_(ツ)_/¯ Ask the mapper who originally mapped color=* there.
You can run an Overpass query with ["leisure"]["colour"] to see if this kind of issue occurs more often, and thanks to me you don't have to check both ["leisure"]["colour"] and ["leisure"]["color"] to get a complete result.
-
Comment from maraf24
What is the colour of this building: https://www.openstreetmap.org/way/373879419/history ?
On what grounds have you decided that it is grey, not white? -
Comment from snoozingnewt
What's better? A less active 'official' discussion, or a more active 'unofficial' discussion? And I find it a bit insulting to generalize all osm discord users like that.
-
Comment from snoozingnewt
And @maraf24
Friendly_Ghost isn't adding any colors... ask the original mapper, not him. He's only changingcolor=* to colour=*. -
Comment from maraf24
@snoozingnewt: That's suprising. You are the second person who thinks I asked about the validity of colour tag on park object, where I specifically asked about the validity of changing the name of the tag. Or in other words - why leisure=parks were not filtered out in the query.
My conclusion is: queries and logic for these changes are flawed and this changeset should be reverted.
If these edits was discussed before on Discord channel than this discussion was also flawed. -
Comment from snoozingnewt
You... were asking about the color of a building. I was responding to that part. Obviously, the original question is something I cannot answer. I don't get your point.
-
Comment from snoozingnewt
"What is the colour of this building: https://www.openstreetmap.org/way/373879419/history ?
On what grounds have you decided that it is grey, not white?" -
Comment from Friendly_Ghost
My apologies to G1asshouse, I'll reply properly to you when I have more time.
@maraf24 Thank you for demonstrating that our map/database contains many different inconsistencies. I cannot fix all issues on my own, so if you think parks shouldn't have a colour, please file a JOSM and iD issue so the validator can pick it up the next time someone tries to edit something similar, and help us by fixing the currently existing "coloured parks" with the overpass query I mentioned.
Concerning that one building you mentioned, it was already tagged as a building with building:colour=grey before I came along. I did not change that and I did not decide anything at all.
-
Comment from Friendly_Ghost
@G1asshouse
> The mailing list is the agreed upon final community discussion location. Just because you've isolated yourself with like-minded people, on discord, does not make it an appropriate alternative to the mailing list.
I find it funny that you think I have “isolated myself with like-minded people”, because it shows your lack of familiarity with the Discord server. You’re very welcome to join it and participate in discussions and chatter.
I’m absolutely not bringing every tag change I make to the mailing lists. What’s the threshold for contacting them? Ten changes, a hundred, a thousand? Getting bogged down in bureaucracy would mean that none of us will have any time left to just map. I’m sure you wouldn’t mail them for a tag change from highway=Residental to highway=residential, I’m doing exactly the same except I’m doing it on a larger scale.
> My point that you do not know if there are OSM editors whom prefer to tag with color instead.
I don’t need to know that. What you, I and all other mappers need to know is that color and colour mean 100% exactly the same, that OSM has a guideline to choose British over American English, that colour is many times as popular as color and that data/map users find our data more useful if it is consistent. I contributed to that in this changeset and others and you're very welcome.
-
Comment from Korgi1
Classic European forcing eurocentric spelling on us poor Americans.
-
Comment from Friendly_Ghost
@Korgi1
1. There's no point in playing the victim. You know as well as everybody else that OSM in general prefers British spelling over American spelling. You can also compare former color=* usage (see this changeset) with https://taginfo.openstreetmap.org/keys/colour#chronology . You'll find the same trend for other keys where GB & US spelling differ.
2. That's not eurocentric, that's a global thing. See https://taginfo.openstreetmap.org/keys/colour#map for reference.
-
Comment from G1asshouse
@Korgi1
(I understood your sarcasm)
Please don't gaslight anyone. It's not about the spelling. -
Comment from G1asshouse
@Friendly_Ghost
> The mailing list is the agreed upon final community discussion location. Just because you've isolated yourself with like-minded people, on discord, does not make it an appropriate alternative to the mailing list.
I find it funny that you think I have “isolated myself with like-minded people”, because it shows your lack of familiarity with the Discord server. You’re very welcome to join it and participate in discussions and chatter.
I’m absolutely not bringing every tag change I make to the mailing lists. What’s the threshold for contacting them? Ten changes, a hundred, a thousand? Getting bogged down in bureaucracy would mean that none of us will have any time left to just map. I’m sure you wouldn’t mail them for a tag change from highway=Residental to highway=residential, I’m doing exactly the same except I’m doing it on a larger scale.
> My point that you do not know if there are OSM editors whom prefer to tag with color instead.
I don’t need to know that. What you, I and all other mappers need to know is that color and colour mean 100% exactly the same, that OSM has a guideline to choose British over American English, that colour is many times as popular as color and that data/map users find our data more useful if it is consistent. I contributed to that in this changeset and others and you're very welcome.“...it shows your lack of familiarity with the Discord…”
“... join it and participate…”
My discord personal history with Discord in general and OSM discord servers in particular is irreverent. Though, it is clear that you don’t know who I am and have no idea if I’m already there or even how long ago I joined. I am aware of you though.“Getting bogged down in bureaucracy would mean that none of us will have any time left to just map.”
The fact that you refer to the larger communities’ request to be included in significant changes to the date base as “bureaucracy” is why your behavior here is objectively grotesque. The whole of the community owns the data set not just you and your friends. To intentionally exclude the community because it inconveniences you is the definition of self-centered. This orientation is obnoxious in life but destructive when in a group setting (such as OSM).“ highway=Residental to highway=residential”
Your “borrowed” example is not analogous to what you’ve done. It’s also irreverent. The complaints are about your actions and not the inclusion or exclusion of the letter “u”. Further complaints are in regard to your behavior to the communities’ complaints.“I don’t need to know that.”
“ What you, I and all other mappers need to know…”
“ you're very welcome”
Honestly, I was going to defend you when another OSM editor called you arrogant. Was. -
Comment from Friendly_Ghost
Dear G1asshouse,
To change the topic back to the original constructive discussion, I am still eager to know how you would approach and resolve the issue that OSM contains a large quantity of redundant and duplicate data.
Best regards
Casper
-
Comment from tekim
> I am still eager to know how you would approach and resolve the issue that OSM contains a large quantity of redundant and duplicate data.
* Write a proposal such as this, and put it on the wiki:
https://wiki.openstreetmap.org/wiki/Mechanical_Edits/malenki
* Send an email to the *official* OSM mailing list announcing your intention, providing a link to your plan, and seeking feedback.
* Incorporate feedback received.
* Execute your plan -
Comment from Friendly_Ghost
Thank you, tekim, for consisely explaining the procedure to me. That looks very reasonable and I'll use that approach for my next large edit.
-
Comment from Lee Carré
This became a soap-opera rather quickly in my absence. Sigh.
If I had the time & motivation, I'd do a point-by-point rebut to each questionable statement, but I feel that such effort would be wasted on someone who presumes to know best.
I notice much buck-passing & denial / rejection of any kind of responsibility-taking on the part of Friendly_Ghost. Along with a recurring theme of ‘not my problem, someone else should figure it out for me’.
Nothing prevents one from making a copy of the DB, with whatever changes, and publishing separately. One could even make a business-model out of supplying OSM data which has gone through some kind of extra processing.
No-one has argued against consistency. The contention is over what's deemed to be canonical, how that's decided, and that it be the collective decision of the whole community.
An attitude of ‘dealing with other people takes too long, so fuck it I'm just gonna make sweeping changes as I see fit anyway’, besides other issues, is how edit wars start (and people get sanctions on their account, like a timed block).
If one has such a compelling case, then it should be easy to persuade others of its merits. To shun this casts doubt & suspicion.
Discord; to avoid the usual lengthy freedom-respecting vs. user-subjugating spiel, I'll keep this blunt; Discord is proprietary, and is at odds with the philosophy behind why OSM came to be. I contribute to OSM because of valuing the principles of libre (especially copyleft) info (data, software, media, whatever). So, expecting me to be a guest on someone else's server, which is set up as a private walled garden, requiring proprietary software and a one-sided contractual agreement (to which I simply can't agree), in order to discuss a libre-data project like OSM … well, that's highly contradictory.
Discord is not open to all. Therefore I don't see how it can be ‘community’. A mailing list, however, is fundamentally different.
But, to preempt some common fallacies (since they seem popular in this discussion), I am not at all against IM-style communications. However, for me, it would have to be a libre system (e.g., IRC, XMPP, Matrix).
There are practical considerations, too; while real-time channels might work well with a group of friends, it doesn't scale well when folks are all in different timezones (unless those in other timezones are also expected to forego sleep, in order to adapt to the demands of individuals who prioritise their own convenience) and who have differing amounts of time to engage (store & forward comms makes it easier to read discussions-of-interest while one was elsewhere, and to then reply selectively to relevant parts).
Mentioning that some private venue is more popular is the epitome of a logical fallacy, and therefore irrelevant.If one's own time is of concern, then learning how to script the process (automation isn't necessarily full-blown software development) would seem in interest. Especially when doing it properly (bounding boxes of sane dimensions) takes 4-6 times longer than the several hours for a singular global changeset.
The time spent learning how to automate is a one-off, but the savings are recurring. Thus, the returns compound (or whatever the inverse of diminishing returns would be).
To rule it impossible (while insisting on continuing with an existing questionable process) because of not knowing how rather begs for the obvious retort; learn.
Otherwise it's akin to saying; sorry, Officer, I don't know how to keep the car driving straight, so me colliding with cars in other lanes is simply unavoidable. I'll give you one guess as to how tolerant said Officer will be to such an attitude of ‘I dunno how, so others will just have to deal with it’. I assure you, he'll find an effective solution to stop your collisions much more rapidly.As for automated edits; there is good reason why they're discouraged. It's easy to make faulty assumptions in one's comfy armchair, compared to facing the reality of what one finds during survey and trying to figure out the least-bad way to encapsulate this in OSMs data model. As someone who lives on a small independent island, which has changed nationalities several times, I oft encounter quite weird head-scratcher situations, which simply don't fit into the anticipated categories of whatever proposal might have made sense for a larger jurisdiction. This is one of the reasons, I imagine, behind textual tagging; it's flexible, to allow for unanticipated cases.
This whole ‘the real-world planet should conform to my idea of how OSM tags should be used’ is backward.
I entirely relate to the desire for a well-ordered dataset, I really do. Problem is, the world is messy, ad-hoc, changed incrementally rather than via grand-plan.
To be dismissive of these truths, is going to alienate others.
To essentially disregard the concerns of others, and stick to one's presupposition, isn't intellectually-honest, or compatible with community. One might even argue that disregarding the community is anti-social, which is at odds with a social project.
I'm reminded of Thomas Sowell; there are no solutions, only trade-offs.
Especially to then ignore the effects of one's own actions, and insist on continuing, and that others do work to make it easier for you … well, hang on, you're the one making the agamgdpg and expecting to continue. How's that anyone else's responsibility? Other people have their own affairs to deal with. If you don't have the time to learn to improve (even if only for your own benefit), then how can you expect others to have copious free time to provide you with the benefit of their learning to be more efficient (at their own cost)?
Give a man a fish, and he'll eat for a day. Teach a man how to catch his own fish and he'll eat for a lifetime.
Why should other engage (let alone indulge or enable) when you're unwilling to (in good faith)?
As happened, it's so much easier to write you off as arrogant. After all, can't reason with unreasonable people.
It's sad to see that this changeset discussion wasn't all that constructive or productive. Especially since the original request (to not have my local changeset list full of irrelevant spam due to enormous bounding boxes of careless / thoughtless changesets) was really quite simple.
My concern for Friendly_Ghost is that he'll have hours of work simply reverted, and that further persistence will yield exponentially-longer blocks.
It is clear that he has enthusiasm, even if it may be misdirected.Might I suggest, Casper, to survey your local area. Fix obvious blunders (e.g. the type of retailer is wrong, or such), add missing features (amenities & small ways) & details (like opening hours, etc.). This would likely be treated as far more valuable. Plus, as surveyor, unless someone else surveyed it around the same time, it would be difficult to argue against the nature of the changes you make (even if they critique your choice of how to tag things).
Maybe once The People's Map rivals the proprietary ones in terms of completeness & up-to-date accuracy, then we can spend hours having academic arguments over the minutiae of tagging (and I say that as one who thinks that colour should be spelled with a U, but that's not the point here).
In this case, re usefulness to dataset-parsers (since they're not consuming the data, but copying & using it in order to produce something else); a bit of RegEx makes the whole colour vs. color entirely a non-issue; /colou?r/
Boom, done; matches both, and then one doesn't have to care which spelling mappers input.
While my next point may sound harsh, I feel it necessary to drive home the lesson; if you knew even a little bit of background re scripting, possibly programming, and other things relevant to automating your workflow, then this would've been apparent to you. Yet, your dismissive refusal to even consider it, and persist with a very manual process … well, it just looks silly.
I say this out of kindness and sympathy. I imagine you didn't intend to appear foolish. Yet, that's part of what likely led others to conclude arrogance; presuming to know best, while clearly oblivious to the basics of the non-problem you're trying to solve (at least with this particular changeset). You also indicated that it took several hours of your time, yet perhaps you can now see how the utility of it is dubious (while adding load to the OSM servers, and increasing the size of the database).Maybe others might know things which you don't. It would be wise to give benefit of doubt and listen (if only in order ten determine for yourself).
In the end, even though it's not their job to explain to you, they're not gonna be willing to try if you seem unreceptive.Anyway. Another advantage of going outside to do surveying is that one has much less time (and is often lacking the inclination) to engage in InterWeb bickering.
If, instead, you wish to continue doing arm-chair mapping, then a would suggest starting small, setting review_requested=yes on your changesets, and learning the prevailing consensus that way.
For an example; I didn't start with iD (I've never used it); I started with StreetComplete (in offline mode, before I even had an OSM account), and paid attention to the changesets it uploaded to see what it was changing and how (peeking at the clockwork behind the face & dials of the pretty UI which only asked simple questions). Yet, I knew I was still contributing positively & constructively, even if I didn't yet know the details.
Then, once I'd groked the basic schema of tagging, I ventured to using editors which didn't hide the details for the cases which SC didn't handle, but still very much sticking to tagging (no adding ways or even nodes) and possibly correcting the position of glaringly mispositioned PoI nodes. All this while still heavily using SC (partly because any tagging mistakes weren't directly my own blunder, but the result of how SC was programmed; I trusted that the developers knew better than me how things should be done (and suspected that this was way it was possible to only change some things, but not everything; maybe those other things had no clear consensus on how it should be done), and later learned that there's quite a lot of consensus-gathering which goes into SC quest-design).
Only as I've become more confident & familiar (being less of a clueless newbie, by starting to correctly figure things out for myself, while still checking the wiki to be sure) did I start adding or modifying ways, adding PoI nodes, and using editors like Vespucci slightly more than SC (when it was clearly more efficient to make many related changes in the same set with Vespucci than the way SC would've done things.Rinse & repeat, for each stage, progressing toward becoming proficient.
Notice the pattern, here. Notice what's conspicuously absent, too.
Making a global change, early on, seemed very naïve. Also arrogant to assume that somehow I had noticed a problem which others hadn't, on such a scale. What was more likely was that such as easy & obvious problem would've already been fixed. Thus, since it exists, there's likely some non-trivial reason why it hasn't been solved of which I'm unaware. Thus, I should leave it alone until I've ceased being clueless newbie for quite a while, rather than diving in because of assuming that the only reason must be because all other contributors are lazy fools.
Especially that this is a global change; two of the core principles of OSM are local knowledge (people contributing data from the area they call home) and a map for the community (rather than controlled by some far-off centralised corporation). To this end, people feeling like they have a sense of local ownership (while recognising & accepting that the data is for everyone), in the sense that they are the authorities of what's in their local environment and thus should be on the map, plus that they have meaningful influence over how the map represents said environment (to account for culture & society, and how they might affect how data is interpreted) that mapping is BY then rather than done TO them, is vitally-important.
So, when someone comes along, from afar, and basically tells them that their way is wrong and this is how it should be done (outside of the technical concerns of a database of tagged nodes; it's not a free-for-all), simply isn't gonna sit well or be taken kindly.I had an example of this, a while ago. I'll keep the specifics vague, because it's not about them, but the general point re social effects.
So, a non-local mapper had added the outlines (from imagery) of a bunch of buildings. Normally this is helpful, and I wouldn't complain. However, they had all been tagged as the same specific type (rather than just building=yes to await a local to do a survey). This collection happened to be within easy walking distance for me at the time. However, when I was there, the assumption of building type / purpose was clearly wrong in several cases. Very wrong (not simply variation on same type, but an entirely different type). Since the tag wasn't building=yes, I couldn't simply use SC (which would've been quick & easy to do the lot of them), and much correction had to be done in a more capable editor, which took more time.
Later, after doing all that, I commented on the original changeset to point out the problem.The contributor who added the outlines said that (from imagery) they looked like a typical set of the type he assumed they were, as was common elsewhere.
And there's my point; yes, fine, elsewhere. But the place he was drawing from is different to my island (yes, mine, because I'm as local as it gets). Quite a few things are different, here.
His presumptive attitude was irritating, especially that correcting the results had taken much more time than had he just set building=yes. I (less bluntly) explained this, too, and the benefits (e.g. SC quests) of not assuming.
I also pointed out how much longer it took, and that there's only 2 locals who do any (recent, regular) surveying (so our survey time is naturally precious).
Had he done the minimum, I would've been delighted, because it was fair amount of work, which would've helped out the handful of locals quite a bit (as has been the case when others have traced a bunch of building outlines). But, the assumptions made my survey take several times longer than otherwise, which made it a problem. Enough of one for me to remark on the changeset.
So, I hope you might see why smaller changesets and working WITH (local) communities is preferable to what amounts to a glorified global search-and-replace sting-manipulation exercise, which for many would've delivered the message of ‘you're mapping your own environment wrong’.
Anyway, that's plenty of food for thought, I'm weary of explaining what I feel is (or should be) obvious, and need a break from this petty squabble for another several days.
-
Comment from Woazboat
Welp, I just spent an hour writing a sizable comment in response only to lose it all because I acidentally clicked on a link while trying to resize the comment text box....
To sum it up:
I agree with the changes by @Friendly_Ghost and putting it into a single changeset. Changing `color` to `colour` is clearly correct with little chance of accidental harm. Having multiple spellings and variants of the same thing is harmful to OSM, leads to more unnecessary work for data consumers and has detrimental effects on the software ecosystem around OSM and the usefulness of the database as a whole. (Correcting for one mistake in software is fairly easy, correcting for hundreds is not. Things that could be fixed with a simple change like this in the upstream database have to be worked around a hundred times over in every single downstream application.)Breaking up large changsets like this without any clearly disjoint clusters does not provide much benefit and has some clear disadvantages. Changesets exist to group changes that belong together. This benefit is lost when the changes are split up and understanding and working with multiple spread out changesets is much harder. Yes, having a changeset like this pop up in your feed when there's no acutal change in your area can be annoying, but the large bounding box cannot really be avoided when the changes actually are spread out across the world. Breaking up the changes into smaller areas does very little to avoid this as there will most likely still be one of these smaller changesets covering a particular area.
-
Comment from drkt
I've said this in the Discord and I'll say it again, here.
Verbatim:
"This changeset size thing just sounds like a poor software problem, to me, not a poor practice problem. I grant my domain is very tiny, but there is a mapper in my country who makes changesets that span my entire domain pretty often, but it's all good data. I use whodidit to see what the changes actually were in the area I care about, rather than review their whole changesets everytime they cross my domain"To add / clarify:
You have the tools available to filter out bboxes that cover your area but add nothing. That you choose not to use them, or are not aware of them, is not their problem. This particular changeset goes over my area, while changing nothing in it. This went completely over my radar because I don't go by bboxes. I literally did not even know this changeset spanned my area.
- 339599686, v4
- 339599687, v4
- 339599688, v4
- 339599689, v4
- 339599690, v4
- 339599691, v4
- 339599692, v4
- 339599693, v4
- 339599694, v4
- 339599695, v4
- 339599696, v4
- 339599697, v4
- 339599698, v4
- 339599699, v4
- 339599700, v4
- 339599701, v4
- 339599702, v4
- 339599703, v4
- 339599704, v4
- 339599705, v4
- BBC Corfu East Road Race Course (73463), v9
- Walderlebnis-Pfad (165234), v4
- Reitweg (165488), v2
- Nordic Walking (165491), v3
- Gansrød folkesti (kort) (280718), v5
- Gansrød folkesti (rullestol) (280720), v6
- 012a (1086365), v11
- Howe Sound Crest Trail (1150682), v10
- Stratum West (1190715), v10
- Gellep-Hafen Ltg 1 (1190717), v4
- Gellep-Hafen Ltg 2 (1190719), v4
- Loipe Waller 1 - Schupfer Schleife (1328761), v6
- Loipe Waller 4 - Grafenbucher Forst (1331963), v6
- Kirkerunden (1340284), v15
- Skøyenhagan (1341486), v6
- 1351808, v4
- Loipe Neutras - Hubmersberg (1360973), v2
- 1514761, v3
- Bus 721 Aubonne, gare -> Féchy, village -> Bougy-Villars, poste -> Aubonne, gare (1524021), v56
- 087 rot (1535619), v10
- 8941535584, v1
- 459977298, v3
- 810119060, v3
- 1314371774, v4
- 1314371780, v5
- 1314371788, v4
- 1314371790, v5
- 1526909108, v2
- 1526909111, v2
- 1526909116, v2
- 1526909117, v2
- Black Josnel shop (1595257506), v7
- Salwagon Temoun Jehova (1596537702), v8
- Bon Boutique (1596627927), v5
- Ti Privelege (1596783103), v4
- College University De Maniche (1596787250), v4
- Shop Eternel R Grand (1597702659), v4
- Radio VPS (1598072783), v6
- Eglise Baptiste De Suzon (1599119735), v5
- Central Michelet Borlette (1692191190), v5
Welcome to OpenStreetMap!
OpenStreetMap is a map of the world, created by people like you and free to use under an open license.
Hosting is supported by Fastly, OSMF corporate members, and other partners.
https://openstreetmap.org/copyright | https://openstreetmap.org |
Copyright OpenStreetMap and contributors, under an open license |