Talk:Microgrants/Microgrants 2020/Proposal/A 360 OpenStreetMap editor
Editor
The idea of integrating street level imagery in a web based editor was previously demoed in Deriviste by Richard, albeit not using 3d imagery. Writing a new editor from scratch is a fairly complex task, which tends to take fairly extensive amounts of time. As an example, Deriviste only supports adding new POIs at this time and refers to well established editors like iD, JOSM,... to clean up the data later on. Can you maybe sketch which use cases you plan to cover by this editor, and which existing code bases you plan to use. Once this project is done, how would you envision supporting this new editor going forward? Mmd (talk) 08:21, 16 May 2020 (UTC)
- Yes, you are right, Deriviste is a similar concept, but its functionality leaves much to be desired. The key difference is that I offer to edit the existing data, not only adding POIs. Of course I do not plan to include all the features of the established editors. I'll start with small things like moving nodes and editing tags, which is a fairly doable task. I have already done more than Deriviste, if you see the links at the beginning of the proposal you can find the example with OSM data overlaid on top of the 360 photos. Speaking of existing code base I'm using Photo Sphere Viewer a framework for displaying 360 photos and drawing geometry on top of it. Use cases that I envision are POI editing and object measurement. For example, it is possible to measure building heights, and of course to add the number of floors. I have plenty of ideas, and I believe other people will add their feature requests too once it starts working. My vision is to make it an open-source project so anyone could contribute to it. SviMik (talk) 11:44, 16 May 2020 (UTC)
Server rent
I didn't quite understand the topic of server rent to host the editor. If I get this right, you would eventually upload imagery to Mapillary, fetch OSM data via the API (rather than from your local server?). So this would boil down to hosting some html/js/css files and everything else would happen in a browser window? In case you're planning to host imagery which other people would upload, a server for 70€/month seems fairly small due to the large amount of disk space required. So what exactly is the purpose of this server? Mmd (talk) 08:26, 16 May 2020 (UTC)
- First of all, there are several algorithms that I plan to run on the server, the most promising is OpenSFM, which potentially would allow me to improve the camera positioning accuracy. I also plan to allow uploading to the server directly as an option, so it wouldn't depend on third-party services like Mapillary. And of course, everything can't happen in the browser window. There are number of features that may require database, image processing, etc. Once the project is running I can decide if the server cost could go up or down and determine the best course of action depending on its popularity and how it will be used. SviMik (talk) 11:44, 16 May 2020 (UTC)