Mapbox's engineering blog on mapping, WebGL rendering, spatial search algorithms, and location features like isochrones.
HN commenters link it for technical posts, such as the WebGL wind map and the R-tree spatial search dive, which one reader found useful.
Awesome to see a new release! We've followed Cockroach for a while, would be a good fit four our workload.
However, I think it's sad that their pricing is so opaque.
Even after a few emails back and forth with their sales people it still feels fairly arbitrary, and since it's not transparent we don't know if they're just making numbers up based on what they think we're willing to pay.
I find this sales-tactic off-putting. In theory it allow them to "not leave any money on the table", as they would probably frame it, but I'm curious how many deals they will lose simply because people don't want to be held hostage, or don't want to put up with the sales-hassle.
With AWS and Google Cloud (their main competitors), we know that prices won't be jacked-up.
———
An interesting example of this mistake is Mapbox.
They had a limited free tier, where most use-cases would be "talk to sales for pricing". We had a frustrating 2 months talking with their sales people, finally getting a reasonable deal. But we were very close to just ditching them for Google Maps (even though I liked Mapbox product better and was willing to pay more for it), just because the price negotiation and all the cheap sales-tactics was so time-consuming.
Then, 6 months later, Mapbox CEO posted this blog post:
https://blog.mapbox.com/new-pricing-46b7c26166e7
[excerpt, from Mapbox CEO announcing pricing/sales rethink]
How we price our tools has a significant impact on how those tools get used
…
1. I knew that I had a lots of time. From a business perspective, this doesn't make perfect sense, I know. I knew C++ (sort of) and was looking for a similar language. Learning Rust took me roughly three weeks (coming from C++), it is not as complex as you think. I think the time saved hunting down null pointers has amortized that investment greatly.
2. Again, I didn't know at that time if that was the fault of Windows. All that I saw was an application that was extremely slow, written in either C++ (with Win32) or .NET (both of which are possible). Second, the application had very shitty PDF output (ex. a shape with CMYK was "emulated" by four shapes drawn on top of each other, one for black, one for magenta, etc. - lead to a completely bloated PDF). I suspected that it was the fault of the framework they were using and I didn't want to end up in a similar situation.
3. Display is actually the #1 bottleneck. You have to understand that in a real map needs to draw roughly 10000 lines, fonts, and symbols. The old program (which I had to work with), did this using Win32 drawing routines. Every time you zoom in / out, it could take 1 - 5 seconds (if lots of things were on the screen). In OpenGL, I can do it in a sub-millisecond time (using this technique -
https://blog.mapbox.com/drawing-antialiased-lines-with-opengl-8766f34192dc paired with primitive restarting in OpenGL (https://www.khronos.org/opengl/wiki/Vertex_Rendering#Primitive_Restart). ArcGIS, for example, has…
Another technique related to voxels is the "Signed Distance Field" (using a data Field or procedural Function).
Signed Distance Function:
https://en.wikipedia.org/wiki/Signed_distance_function
2D Signed Distance Fields are used for rendering fonts and other shapes defined by a texture whose signed pixels (-128..128) specify how far each is from the border (+ is outside, 0 is on the border, - is inside). They scale beautifully because you can smoothly interpolate between samples.
Drawing Text with Signed Distance Fields in Mapbox GL:
https://blog.mapbox.com/drawing-text-with-signed-distance-fields-in-mapbox-gl-b0933af6f817
TextMesh Pro - Font Asset Creation Process:
https://www.youtube.com/watch?v=eXhfygRgIVY
You can convert them to piecewise outline contours (like an iso-map slicing at many different elevations) with the Marching Squares algorithm, but it's much more efficient to render them directly pixel-by-pixel with a shader.
Marching Squares:
https://en.wikipedia.org/wiki/Marching_squares
Signed Distance Fields are excellent for shaders efficiently antialiasing and drawing effects like outlines, drop shadows, embossed and beveled edges, glows, multi texture mapping, bump mapping, etc). The shader can quickly determine exactly how far away each pixel is from the outline, and blend the appropriate effects together for that pixel, so you get precise anti-aliasing, outlines, drop shadows, etc, practically for free! (And literally for free, now that Unity gives…
Back when I was at Mapbox and we were developing the library used for this sort of visualization, I did a take on my home city of Portland, OR in a similar fashion:
https://blog.mapbox.com/visualizing-an-entire-citys-buildings-live-with-runtime-styling-453fe7e39ae6
(Disclaimer: some of the images, and especially the code formatting, don’t seem to have survived very well)
I wrote there a bit more about the data processing side of this, and I tried to make a compelling mobile visualization that was fast. I’d use it a lot for demo and pitch-type meetings, and especially with an HDMI connector direct from iPhone to a 50, 60 or 70” display, the frame rate and fluidity was super impressive.
To this day, I just plain love the tech that we built.
>Yeah, AR is going to find it’s first killer application in the next year or two.
My money is on something like Nexar [0]. AR enhanced driving is going to save commercial fleet operators incredible amounts of money while we're still reliant on human drivers. Real level 3+ autonomy is still years away, but AR HUDs are going to be huge very soon, and the mapping data they collect will build the basis for those autonomous systems. MapBox just released their new Vision SDK that does deep learning based AR locally on mobile devices that is going to have some amazing applications too [1].
[0] https://www.getnexar.com
[1] https://blog.mapbox.com/vision-sdk-d47486656821
I would guess that it's "A demonstration of all the 3D data already in OpenSourceMap"
As heights and number of levels on the building has been a part of the dataset for a while now:
https://wiki.openstreetmap.org/wiki/Key:building:levels
https://wiki.openstreetmap.org/wiki/Key:height
https://blog.mapbox.com/mapping-3d-building-features-in-openstreetmap-7685ee12712a
It's obviously not complete. I took 2 locations that I know well (different countries) and the first has relatively realistic looking 3-d building shapes. The other has nothing.
And that it's using a 3rd party engine. Since the OSM data is "open", there's nothing wrong with making use of it.