I'm slowly working on my manifesto for the #OSMF election, and writing my answers to the suggestion questions. I'll let yous know when it's done
@karlos yes I would support that. 🙂 (that's one of the official questions, so there's more written there too)
I gave a talk at SotM about porting the osm-carto style to vector tiles. I have tech experience with this.
Alas vector tile tech is often a PITA, and I think that'll be the main problem 😔
@karlos there are many design decisions when it comes to vector tiles, which affect things. Like do you require only js/client side rendering? Do you redo the style or port OSM carto? Etc
@karlos I really like RichardF's idea for vector tiles. Ah someday....
@rory Looks like I will get lost if i start to read about topic Vector Tiles. To many experiments, arguments, discussion paces. As a first step, this needs organized.
Even the topic "Vector Tiles" are not clear. To me it is files with vector data. An (web-) app with individual styles will render, may be even in 3D.
* I can't find "vector" on the (official) questions page.
* I found your speak: 2016. Well, I would start with a style-shape mimic Carto an let the user make variations.
* JS? Yes, but not only. WASM of course. A renderer on Git with C++ or RUST to use anywhere.
* More but client? Hm, a server rendering bitmaps with all the advantages of vector: seemless zooms, switching styles, filtering POIs, ...
RichardF? You mean github.com/systemed/tilemaker - yes, we need one/some "standards" what tiles will contain (and how coded). May be additional tiles, if you need 3D data i.E.
I would like to help, if a project exists.
Except "I may be to old for this"
@karlos alas, there just isn't a lot of rendering software and it's a PITA. There's mapnik, mapboxgl, tangram, ish. There's just a lot of lacking software, and companies like Mapbox have moved on & Mapzen are gone.
@rory Lacking software? Web search about OSM vector render software seems to take days. Most of it is commercial and I am not amused about Mapbox either.
* Tangram looks good. There is also C++ usable with WASM for a (Progressive) Web APP
* Mapnik is C++ and Python. Hm?
Why not use one of them?
I wouldn't mind to start from scratch and just use the logic of Carto, not the code.
What would you like, C++ or Rust?
I wrote OSMgo.org as a 3D renderer, just a small part of what we need of course.
@karlos So I don't want to damp your eagerness. If you wanna make something, then please do it. vectortiles on OSM.org would be _great_!
@rory 🤓I started to learn/use Rust and WASM. The idea is to code an example, how OSM vector rendering could be done.
Could OSM found a taskforce to list needed steps and work out ways to solve them?
@rory 😒 I try to do
but the call from JS to WASM
does not work.
I need a running example ...
@karlos I've been programming rust, but I've never done any wasm stuff I'm afriad... I dunno.
The MVT format is very simple. It's just a simple protobuf format. That's not really the hard part. IME it's doing all the graphics & rendering & stuff. I've half started something which renders MVTs as image using cairo (which is what mapnik uses too).
@rory I mostly agree with RichardF. I am puzzled about his problem with carto style. Of course there must not be a hard-coded carto style! We need a style syntax at start. The carto vector style will just be a prove of concept.
I did not have a close look into the vector tile file format. I hope it is something like open topo map to avoid redundant data. My dream is to get tiles by an OSM.org-API. Processed by TileMaker. Including All the validating of OSM tagging and putting into GIT layers.
The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!