Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
>> I'll be curious to hear feedback from this, too. Thanks for your efforts, >> Lukas: I genuinely hope they help our map! > > I can see a use when you have three consecutive segments of a road, where the > first and the last are named (the same) and the middle is not. This might > indicate an omission. > [...] > Every unnamed way that branches or has a junction is difficult to name in an > armchair way. This is exactly what is specified in Lukas’ repository (see https://gitlab.com/ltog/ohni#what-is-this-software-about): "most corner cases are rejected”. I think the tool caters to a mapper which is rather proficient with the tools, as one has to download a .pbf extract and massage the data with Lukas’ tool. This is not something for armchair mappers which blindly click on a green checkmark in Osmose. Bravo Lukas, e cooli Sach! Greetings from Switzerland, Habi signature.asc Description: Message signed with OpenPGP ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
I am a bit unsure on how to proceed from here... Maybe let's start with some technical facts: > IMHO a highway=service, track etc typically do not have names. > Hopefully you account for that. > And i am not shure the assumption that other road classes > always have names holds up. In the current incarnation, in order to show up, a nameless way must be connected to two other ways that have exactly the same tags and also all three ways must be of the same higway=*. > Higher class roads may have ref= but not name=* Currently I am only considering interpolating missing name=*, but looking at gaps of ref=* might be a good idea for a future version. > I use Geofabrik's excellent OSM Inspector Yeah, nice! I wrote the "Addresses" view of it. (code is at https://github.com/ltog/osmi-addresses ) What I do not like so much about it: it shows you a lot of errors, but there is not much you can do about it after you fixed your reachable neighbourhood. It's the same for the "Highways" view (options "Highway name/ref") that looks at the same issue as the "ohni" software: https://tools.geofabrik.de/osmi/?view=highways=8.00885=47.99930=9=Geofabrik%20Standard=name_missing_major%2Cname_missing_minor I specifically wanted the software to give the users the possibility to find spots where they can not only watch, but contribute with a low number of false-positives. I was hoping for feedback especially on that aspect (trade-off between false-positives and high number of results). > if you have a road in 3 segments A B C and A+C have the same name, > then not only does it seem reasonable to me to add the name on B > but also the reply "do a survey" is a dogmatic answer: the ground > does not contain a sign every time osm cuts a way because > of a change in the number of strips for example. > So by survey, you will be reduced to deducing that a segment between > 2 others with the same name, probably also has the same name That is also my point of view. Best regards Lukas ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
> I can't really think what the Osmose people (in this example) could do to > make people > NOT blindly make changes in this way > Add ability to ban accounts from using Osmose? And use it to ban people misusing it? In theory there can be also some sort of fake elements clearly labelled as "mark as false positive if you actually review reported errors" and ban people who will ignore it and blindly click "proceed with edit". (not saying that either is worth doing, not sure how large is the problem and whether either would substantially help) Maybe just reporting abuse of Osmose to DWG for zero-hour blocks (or longer for repeated offenders) is the best solution. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
On Sun, Nov 27, 2022 at 09:16:14PM +0100, Lukas Toggenburger via talk wrote: > Hi all > > As you might know, OSM data contains a lot of highway=* without > name=*. Check your region using the following query: > IMHO a highway=service, track etc typically do not have names. Hopefully you account for that. And i am not shure the assumption that other road classes always have names holds up. Higher class roads may have ref= but not name=* IMHO any assumption about name= (On any object) is pretty difficult and not worth the hassle. Flo -- Florian Lohoff f...@zz.de Any sufficiently advanced technology is indistinguishable from magic. signature.asc Description: PGP signature ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
On Nov 28, 2022, at 5:57 AM, Maarten Deen wrote: > Your remark seems reasonable ;) Thanks, Maarten, I’m chuckling with mirthful laughter here. > Thing is: this is not meant as a bot, so the usual caveats apply. That IS an important consideration; thanks for highlighting that aspect. I didn’t know there were / are “usual caveats,” and if there are some, they should be widely known, especially by people who discuss these things in a place, manner and time such as “this, here and now." > It just serves as a highlight of "something might be wrong here", like so > many QA tools do. What the user wielding the QA tool does with that is his > choice. > Does he automatically correct it? Wrong for OSM standards, but who is going > to stop him. Just like who is stopping anyone using a QA tool and > armchairmapping something that he really can not see from a distance. I’m aware of this distinction and again, it is important. Among others (OSMCha [1], there are any number of such tools...), I use Geofabrik's excellent OSM Inspector [2], which “merely shows,” (and superbly) rather than “and fixes, too." Perhaps another example will help: JOSM’s Validator tool gives the opportunity (between clicking the Upload button and actual data being uploaded to OSM) to review “flagged” problems, some as mild Warnings which could be ignored (but shouldn’t be), some as more serious Errors which really ought to be solved. For some of these problems, JOSM is clever enough to “light up” (activate in its user interface) a “Fix” button which is smart enough to “take the right OSM actions” to actually fix the problem. This is great when available, as it solves the tedium of manually doing something which can be automated (so, click the Fix button). And, as JOSM (and OSM) develop, while more and more identified problems are automatically-solvable, some certainly are not, and so the Fix button remains dimmed, meaning “these must be fixed manually.” That’s the distinction I’m making here: I haven’t analyzed Lukas’ code to see where / when / whether (and how) he does this (and again, there are certainly cases where this MIGHT be possible, and where it is, great, do so). I am simply saying “there are times and places where an automated tool CAN fix things” (like naming nameless highways…though this really IS tricky, speaking from personal experience), and there are absolutely times and places where it can’t. Both tool developers (especially) and also, those who wield such tools must be aware of this “sometimes we can, sometimes we can’t” distinctions. And I’m not saying Lukas has done this, but in this realm (and because I know a couple things about “quality” and “mapping” I can say this): it is all-to-easy to glibly make assumptions which are better left not made. Especially in the context of OSM, wider vetting (exactly what Lukas is asking for here) is exactly what is needed. Though, as we get wider input that includes “seems reasonable,” I urge caution. I’ll stop here, as I don’t want to repeat myself. [1] https:osmcha.org [2] https://tools.geofabrik.de ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
On 28/11/2022 13:57, Maarten Deen wrote: What the user wielding the QA tool does with that is his choice. Indeed, but as we've seen in lots of places users sometimes blindly follow "suggestions" without engaging their brains. As an example, Osmose contains very clear information that users should not blindly obey it (see https://osmose.openstreetmap.fr/en/map/#zoom=0=39.9=-74.6 and https://wiki.openstreetmap.org/wiki/Osmose ) yet there are still many complaints about people using Osmose for "fixing" stuff in OSM that isn't actually broken. I can't really think what the Osmose people (in this example) could do to make people NOT blindly make changes in this way, and it's really useful to the project to have something suggesting what might (but might not) be a problem**. Perhaps what needs to happen is for people who actually do go outside and map stuff to push back more forthrightly and more often, when we know that something is wrong, and even when it "just looks a bit iffy". Best Regards, Andy ** for the avoidance of any doubt I'm not complaining about Osmose here! ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
Your remark seems reasonable ;) Thing is: this is not meant as a bot, so the usual caveats apply. It just serves as a highlight of "something might be wrong here", like so many QA tools do. What the user wielding the QA tool does with that is his choice. Does he automatically correct it? Wrong for OSM standards, but who is going to stop him. Just like who is stopping anyone using a QA tool and armchairmapping something that he really can not see from a distance. Regards, Maarten > Op 28-11-2022 14:15 CET schreef stevea : > > > See, saying “seems reasonable” actually seems reasonable, until one realizes > one doesn’t truly know. Ask yourself if others in OSM would agree if “seems > reasonable” is good enough to meet OSM’s criteria for data entry: you’ll get > mixed answers, though a sizable number will say “not really good enough.” > You might even have a very high degree of confidence…though, ask yourself if > you want to navigate (or otherwise rely upon) a map with what amounts to > guesswork. That’s how the camel’s nose (of creeping errors, one datum at a > time) gets into the tent (map). I mean no disrespect to camels. > > I have decades of experience in software quality assurance at top companies > (Apple, Adobe…), so I have great respect for Lukas’ tool finding / > identifying errors (emphasis on those verbs), it’s what is done after that > which matters. Guesswork? Mmm, no, I’d prefer not. Our usual “on the > ground verify” (or otherwise equivalent, like “I already know that”) > criteria: yes, much better. > > We’re not quibbling (slightly objecting to trivial matters) here: these are > fundamental decisions each and every mapper makes as they enter data into our > map database. I strive to keep that quality as high as I possibly can, > though everything I say here is simply one person’t opinion. Let’s be > careful with power tools: they’re great at finding / identifying errors, > whether they can “fix” the data after that must be carefully considered > case-by-case. > > > On Nov 28, 2022, at 12:44 AM, Marc_marc wrote: > > Le 28.11.22 à 00:43, Dave F via talk a écrit : > >> a "high confidence" interpolation, from an armchair or anywhere, will lead > >> to inaccurate data being added to the OSM database. > > > > if you have a road in 3 segments A B C and A+C have the ssame name, > > then not only does it seem reasonable to me to add the name on B > > but also the reply "do a survey" is a dogmatic answer: the ground > > does not contain a sign every time osm cuts a way because of a change > > in the number of strips for example. > > So by survey, you will be reduced to deducing that a segment between > > 2 others with the same name, probably also has the same name > > > ___ > talk mailing list > talk@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
See, saying “seems reasonable” actually seems reasonable, until one realizes one doesn’t truly know. Ask yourself if others in OSM would agree if “seems reasonable” is good enough to meet OSM’s criteria for data entry: you’ll get mixed answers, though a sizable number will say “not really good enough.” You might even have a very high degree of confidence…though, ask yourself if you want to navigate (or otherwise rely upon) a map with what amounts to guesswork. That’s how the camel’s nose (of creeping errors, one datum at a time) gets into the tent (map). I mean no disrespect to camels. I have decades of experience in software quality assurance at top companies (Apple, Adobe…), so I have great respect for Lukas’ tool finding / identifying errors (emphasis on those verbs), it’s what is done after that which matters. Guesswork? Mmm, no, I’d prefer not. Our usual “on the ground verify” (or otherwise equivalent, like “I already know that”) criteria: yes, much better. We’re not quibbling (slightly objecting to trivial matters) here: these are fundamental decisions each and every mapper makes as they enter data into our map database. I strive to keep that quality as high as I possibly can, though everything I say here is simply one person’t opinion. Let’s be careful with power tools: they’re great at finding / identifying errors, whether they can “fix” the data after that must be carefully considered case-by-case. On Nov 28, 2022, at 12:44 AM, Marc_marc wrote: > Le 28.11.22 à 00:43, Dave F via talk a écrit : >> a "high confidence" interpolation, from an armchair or anywhere, will lead >> to inaccurate data being added to the OSM database. > > if you have a road in 3 segments A B C and A+C have the ssame name, > then not only does it seem reasonable to me to add the name on B > but also the reply "do a survey" is a dogmatic answer: the ground > does not contain a sign every time osm cuts a way because of a change > in the number of strips for example. > So by survey, you will be reduced to deducing that a segment between > 2 others with the same name, probably also has the same name ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
Le 28.11.22 à 00:43, Dave F via talk a écrit : a "high confidence" interpolation, from an armchair or anywhere, will lead to inaccurate data being added to the OSM database. if you have a road in 3 segments A B C and A+C have the ssame name, then not only does it seem reasonable to me to add the name on B but also the reply "do a survey" is a dogmatic answer: the ground does not contain a sign every time osm cuts a way because of a change in the number of strips for example. So by survey, you will be reduced to deducing that a segment between 2 others with the same name, probably also has the same name ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
Hi all Thank you for your feedback so far. > I can see a use when you have three consecutive segments > of a road, where the first and the last are named (the same) > and the middle is not. This might indicate an omission. This is what I do. See https://gitlab.com/ltog/ohni/-/blob/main/README.md#user-content-what-is-this-software-about > Any comparison has to be done against an authoritative database > or on ground surveying, for the area in which you're searching. If you try out the software you will find that in most cases quite short segments of highways are reported (often tunnels and bridges) where the name probably was added later on but was missed on that small segment. Of course common sense should still apply when mapping, meaning e.g. skipping questionable cases (described as step 13 in my example workflow). And I definitely do not advocate using this in an automated fashion. Best regards Lukas ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
> Op 28-11-2022 05:18 CET schreef stevea : > I'll be curious to hear feedback from this, too. Thanks for your efforts, > Lukas: I genuinely hope they help our map! I can see a use when you have three consecutive segments of a road, where the first and the last are named (the same) and the middle is not. This might indicate an omission. But it is not guaranteed of course. Every unnamed way that branches or has a junction is difficult to name in an armchair way. Regards, Maarten > > > > On Nov 27, 2022, at 3:43 PM, Dave F via talk wrote: > > > > Most roads don't have names. > > > > Any comparison has to be done against an authoritative database or on > > ground surveying, for the area in which you're searching. > > > > "where the name can be interpolated from neighbouring ways. This allows to > > detect and armchair-fix a (small) subset of these cases with high > > confidence. " > > > > I have a "high confidence" interpolation, from an armchair or anywhere, > > will lead to inaccurate data being added to the OSM database. > > > > Cheers > > DaveF > > > > > > On 27/11/2022 20:16, Lukas Toggenburger via talk wrote: > >> Hi all > >> > >> As you might know, OSM data contains a lot of highway=* without name=*. > >> Check your region using the following query: > >> > >> https://overpass-turbo.eu/?Q=way%0A%20%20%5Bhighway%5D%5B!name%5D%0A%20%20(%7B%7Bbbox%7D%7D)%3B%0Aout%20body%3B%0A%3E%3B%0Aout%20skel%20qt%3B > >> > >> I wrote a Python tool (using Sarah Hoffmann's pyosmium) at > >> https://gitlab.com/ltog/ohni that is able to detect such highways in a > >> planet (extract) file and report the ones, where the name can be > >> interpolated from neighbouring ways. This allows to detect and > >> armchair-fix a (small) subset of these cases with high confidence. The > >> tool is tailored to minimize false-positives. > >> > >> Please check it out and give feedback. > >> > >> Best regards > >> > >> Lukas ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
Without wishing to "diss" (disparage) Lukas's tool (I haven't evaluated it), I would also urge caution here, for exactly the reasons DaveF outlines. I'm also a partly-armchair mapper, but not (usually, if ever) using Python tools, rather knowing that my "armchair-ing" is going to be of high quality because I well-understand (and have many years of practice) with OSM's tenets of "on the ground verifiability" and such. I'm also a "real world" mapper (where I "get out into the real world" and use a GPS and a notepad/pencil...), which I believe should be a prerequisite for being an armchair mapper: that's how you learn and get good at knowing how to armchair-map with high quality and without guesswork that can (and usually does) lead to errors being entered. Especially with going from "no name" to setting a name=* tag to "something" (authoritative), this can be a very ticklish undertaking, I've experienced the difficulty in doing this first-hand, and I usually "duck out" (end the endeavor) as "too likely to introduce specious errors into our map data." High quality armchair mapping (which does not introduce errors) is not easy: it takes practice, knowing what you are doing, likely some deeper knowledge of the geographic area, "how things are done (and/or mapped) around there" and probably something like "I know quite well how to map bike routes (train routes, landuse, forest boundaries..., or whatever you might be mapping)." If you meet all of those "high bar" quality standards, AND you understand what Lukas' Python / pyosmium software does / will do, you might want to check it out and see if it can be a "power tool" for your armchair mapping. I've set high-quality standards for myself (really, I wouldn't map in OSM if I did not), perhaps you should, too. And then, and only then, maybe use power tools to help you, going slow at first, with caution and evaluating your own feedback from the map. I'll be curious to hear feedback from this, too. Thanks for your efforts, Lukas: I genuinely hope they help our map! > On Nov 27, 2022, at 3:43 PM, Dave F via talk wrote: > > Most roads don't have names. > > Any comparison has to be done against an authoritative database or on ground > surveying, for the area in which you're searching. > > "where the name can be interpolated from neighbouring ways. This allows to > detect and armchair-fix a (small) subset of these cases with high confidence. > " > > I have a "high confidence" interpolation, from an armchair or anywhere, will > lead to inaccurate data being added to the OSM database. > > Cheers > DaveF > > > On 27/11/2022 20:16, Lukas Toggenburger via talk wrote: >> Hi all >> >> As you might know, OSM data contains a lot of highway=* without name=*. >> Check your region using the following query: >> >> https://overpass-turbo.eu/?Q=way%0A%20%20%5Bhighway%5D%5B!name%5D%0A%20%20(%7B%7Bbbox%7D%7D)%3B%0Aout%20body%3B%0A%3E%3B%0Aout%20skel%20qt%3B >> >> I wrote a Python tool (using Sarah Hoffmann's pyosmium) at >> https://gitlab.com/ltog/ohni that is able to detect such highways in a >> planet (extract) file and report the ones, where the name can be >> interpolated from neighbouring ways. This allows to detect and armchair-fix >> a (small) subset of these cases with high confidence. The tool is tailored >> to minimize false-positives. >> >> Please check it out and give feedback. >> >> Best regards >> >> Lukas >> >> ___ >> talk mailing list >> talk@openstreetmap.org >> https://lists.openstreetmap.org/listinfo/talk > > > ___ > talk mailing list > talk@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] QA tool for finding nameless highways that are armchair-fixable
Most roads don't have names. Any comparison has to be done against an authoritative database or on ground surveying, for the area in which you're searching. "where the name can be interpolated from neighbouring ways. This allows to detect and armchair-fix a (small) subset of these cases with high confidence. " I have a "high confidence" interpolation, from an armchair or anywhere, will lead to inaccurate data being added to the OSM database. Cheers DaveF On 27/11/2022 20:16, Lukas Toggenburger via talk wrote: Hi all As you might know, OSM data contains a lot of highway=* without name=*. Check your region using the following query: https://overpass-turbo.eu/?Q=way%0A%20%20%5Bhighway%5D%5B!name%5D%0A%20%20(%7B%7Bbbox%7D%7D)%3B%0Aout%20body%3B%0A%3E%3B%0Aout%20skel%20qt%3B I wrote a Python tool (using Sarah Hoffmann's pyosmium) at https://gitlab.com/ltog/ohni that is able to detect such highways in a planet (extract) file and report the ones, where the name can be interpolated from neighbouring ways. This allows to detect and armchair-fix a (small) subset of these cases with high confidence. The tool is tailored to minimize false-positives. Please check it out and give feedback. Best regards Lukas ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
[OSM-talk] QA tool for finding nameless highways that are armchair-fixable
Hi all As you might know, OSM data contains a lot of highway=* without name=*. Check your region using the following query: https://overpass-turbo.eu/?Q=way%0A%20%20%5Bhighway%5D%5B!name%5D%0A%20%20(%7B%7Bbbox%7D%7D)%3B%0Aout%20body%3B%0A%3E%3B%0Aout%20skel%20qt%3B I wrote a Python tool (using Sarah Hoffmann's pyosmium) at https://gitlab.com/ltog/ohni that is able to detect such highways in a planet (extract) file and report the ones, where the name can be interpolated from neighbouring ways. This allows to detect and armchair-fix a (small) subset of these cases with high confidence. The tool is tailored to minimize false-positives. Please check it out and give feedback. Best regards Lukas ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
