fghj753's Comments
| Changeset | When | Comment |
|---|---|---|
| 168734627 | Hi, you have changed this pedestrian crossing traffic light to be regular traffic light (traffic_signals=*) while i couldn't reliably verify which type of signals those are, i'd like to point out that you have also reversed the direction signals are facing to (traffic_signals:direction=*). I assume this is purely accidental change as everydoor doesn't have any indicator to help guess, which way the underlying way is drawn. PS. Since i don't remember traffic lights being here, does this crossing really have signals now? |
|
| 149331125 | Tere, ega sa ei mäleta, kas node/11772485033 on tõesti fooridega ülekäik? Ma pole sealtkaudu ammu liikunud, aga enda teada ma ei mäleta et seal nagu foorid oleks. Küsin sellepärast, kuna üks kasutaja tegi kahtlase muudatuse, et foor paikneb peale ülekäifurada. |
|
| 168733197 | Hi, welcome to OpenStreetMap. Are you absolutely sure this building's walls are made of metal, because that's what building:material=metal means. You may have wanted to add roof:material=metal, which means roof is made of metal.
|
|
| 167972832 | Hi, Please make sure you provide explanation to changesets. This one is spanning 3 countries due to tiny change in way/648202691, while majority of changes seem to be adding buildings near Peterburg. |
|
| 167777512 | Difference from yesterday's changeset/167754914 is that this time i included nodes that were not part of any crossing-way. Overpass query: [out:json][timeout:25];
// Step 3: Remove kerbs that are part of those roads
// Step 6: Intersect with nonroad_kerbs
|
|
| 167754914 | For future reference, here’s the Overpass QL I used to select those kerb=* nodes. First, it limits to nodes tagged only with kerb=* (so we don’t touch kerb tags used alongside bus stops or platforms) - this follows the advice in wiki (kerb=*#By_itself). Then, since tagging a node as barrier=kerb implies it’s an obstacle along a way, we explicitly exclude any kerb nodes that lie on motor-vehicle roads (i.e. all highway=* except footway, cycleway, path), because according to wiki standalone kerb=* may have denoted informal crossing. [out:json][timeout:25];
|
|
| 126353773 | Pretty sure there was bicycle parking at one entrance, but not at another. Bit further away from entrance than cart storage.
|
|
| 85905341 | Hi,
|
|
| 127744153 | For future reference, linking the block message because it contains links to some of Aulo's (user id 417306) changesets and their discussions:
|
|
| 127744153 | Thanks. I wanted to report Aulo's actions to talk-ee month ago, but got overloaded by non-OSM-related stuff.
|
|
| 126138420 | P.P.S @waterproof_19m, ma ei olegi varem EHRist geomeetria importimisele mõelnud, paistab, et nad kasutavad hoone kujude jaoks teistsugust allikat kui Maa-amet. (või on ministeeriumite ühishoone geomeetria registris lihtsalt vigane) |
|
| 126138420 | Warning: very long message. You can just copy this message to Word to make reading it more bearable. Tere, First off, let's start with a brief overview of history. When i became OSM mapper, there was an active building import by SviMik that was focusing on adding new buildings from Maa-amet (MA from now on). SviMik had a policy, that to avoid possible editing conflict (conflict as overwriting someone's work without their consent), his bot would only either add new buildings or update addresses on buildings last edited by selected users/bots (mostly large volume users such as Verbatium, xybot and juhanjuku). At same time buildings added by his MA imports had very high data quality. Because of trying to not interfere with possible future imports, I'm still not really proficient at mapping buildings. Collection of some resources regarding to SviMik's import.
Importing new buildings as outlines is not really difficult per se, just download ETAK data shapefile from MA and upload them with Josm. Problem here is the OSM's expectations for quality. For example, the most obvious step is adding tags that are suitable for OSM object and not uploading buildings that are already added to OSM. However, there are more nuances which reveal themselves after you have decided to start this process. In Estonian context, MA's geometry data is sometimes too high quality for modern OSM expectations (e.g having 40+ nodes per circular way while OSM standard is 20) or has minor artefacts caused by technical limitations of their internal software (complex issue, but it results in too many nodes on a straight line). Those were just issues i had with geometry import. Another dimension is getting the address data right. My import focused only on geometry, because tags were already updated by SviMik's import in 2019. However, i noticed that address field in the dataset was just single text field - it would be really difficult to try to generate address tags from this field. More reliable source might be MA's ADS (aadressiandmed), but since that's not downloadable, just an API, then i believe using it would quickly become bottleneck for the import process. SviMik's source code for import bots is not public, mine is indeed on github repo linked above.
My import on wiki - osm.wiki/Maa-amet_building_geometry_update
Replying to qqqqqqqqqqqqqqqqqqqq's questions:
P.S I was told this comment should be a wiki page. Problem is that I'm replying to your discussion too directly, so converting it to wiki page will require lot of editing. |
|
| 82460204 | Once upon a time i was taught that cyclist can't ride on the street if there is designated cycleway next to it. Double checked traffic laws and streetlevel imagery - Yes, those restrictions can be removed. |
|
| 117835146 | PPS. You have previously mentioned you don't write neither reasonable changeset comments nor split changesets into smaller regions because you don't have time. However, this changeset is part of series of 17 changesets made in 2 hours. I think making two changesets, one for eastern and other for western hemisphere and including adequate changeset comments could have actually been faster than your current process. |
|
| 117835146 | I'd also like to point out language use of this changeset comment. This appears to be global changeset made by russian-speaking user (see created_by value) making changes mostly in Russia (see modified objects list), yet comment on this chset is in German. Dear iWowik, what is this changeset's connection to Germany? Why is changeset comment in German, not English or Russian? PS. According to https://resultmaps.neis-one.org/osm-discussion-comments?user=iWowik , the user has been notified and confirmed receiving the information about large changesets and changesets comments since April 2021. |
|
| 117768233 | Tere, Priit.
|
|
| 116138570 | @SimonPoole That's pretty much default behaviour in Josm - if you use click-and-drag method for selection tool, then it will select all nodes and ways within area by default.
|
|
| 115769750 | Dear Mr/Mrs McBlyat, Judging by the lack of any further complaints I presume reboot has resolved the issue. All the best, Fghj753 |
|
| 115769750 | Dear Mr/Mrs McBlyat, I'm sorry to hear about difficulties you are experiencing with using changeset visualization tools. That software can be highly overwhelming for a person who has just joined the OSM community. Sadly I was not able to replicate issue you were describing on some of the most common tools.
Best regards, Fghj753 |
|
| 115775114 | Nicely done, welcome to OSM!
|