Map adventures in weird web standards - gyroscopes, texture cubes, and mutants

@IvanSanchez gave a presentation on why he believes it would be simpler for Google to share its map tiles than to develop a standard for maps in HTML.

Please comment here in response, I’m certain Iván would appreciate your views.

A video of the presentation is here

Native implementations of web browsers have tended to progress more low-level in recent years. This is not surprising given the business models of the vendors that provide browsers, with the exception of Firefox, and it’s hard to change that. So, there are many points in his presentation that I can agree with.

On another hand, I think that the design of a browser API for various javascript enabled streaming in video element could be very informative for map element as well.

I’ve posted some of the factors below.

Well, the presentation covered two more topics (joystick interaction/accesibility, and capabilities of low-level GL graphics). Those roughly correspond to parts 2 and 4 of my essay.

I want to write down something that I noted quickly on the Gitter chat during the workshop: I think the position of Google in regards to maps standards is hypocrite:

  • On one hand, Ed Parsons confirmed (during the talk, via gitter chat) that loading GMaps tiles in a JS library through a URL template (i.e. L.tileLayer("{x}&y={y}&z={z}&s=Ga")) goes against Terms of Service, (which is what prompted the horrible horrible hacks of GoogleMutant)
  • And on the other, representatives from commercial map providers (including Google) asked to «make things easy».

My immediate, passionate, aggresive gut reaction is to yell «if you want things to be easy, make things easy yourself». Like, start by enabling the GMaps Tile API to all users instead of «select partners», and then wrap that API into a WMTS by doing trivial rewrites of URLs. You know, lead by example for once.

As of now, I have the following strict belief: Google Maps (or any of its representatives) has no moral authority to ask for standardization or simplification of map APIs.

1 Like

So let’s be clear. Google Maps is not asking for standardization of maps. The Maps for HTML Community Group is.

That Ed even takes the time to listen and encourage the discussion is a start.

Finally, keeping things simple does not mean keeping the status quo. Even the original Leaflet library author has told us that simplicity will save GIS.

And you can’t get “simpler” than HTML; it’s the next logical step in Vlad’s hierarchy of productive environments.

I don’t know if I’m in the minority or not here, but I fundamentally agree with Ivan’s messages, especially around Google’s tiles. Let’s face it, they key thing users/developers want that they have is the cartography: it is not only extremely well-done (especially in the “satellite with labels” tileset), but users are so familiar with their stuff, it’s almost a “standard”. Of course Mapbox and several other commercial providers make beautiful maps, too, but Google has a long lead.

Of course, I recognize that Google also has a business model to protect. So what if they published a version of their tile sets that could be used outside of their API? Maybe it lags a few months in updates, has less (or more?) “commercialized” data points, has different rate limits, or something like that. They get to keep some of the fundamental value-added pieces (geocoding, place search) inside their API and still gain some good will.

I’m not sure if I buy that a new MapML standard will really help the web. I think I probably have a bias because I don’t find existing mapping libraries “too hard” to work with. There are so many more choices today than there were in ~2006 when I first started making slippy maps, and I think that’s a good thing.

1 Like

Re: “too hard”, I don’t think anyone says these libraries are that. They’re pretty good from that POV. My concern on that score really is more like: what do we as a global society teach to kids? If you say “Leaflet”, there are others who say “OpenLayers”, or “Google Maps”, or “ArcGIS Online”. Which are all great, as you acknowledge (in the case of Google Maps). Why shouldn’t all of those things be accessible through the Web in a simple (not-Turing-complete, rule-of-least-power) format like HTML? I think the answer is given in my position statement.

There are things in / taken care of / standardized by the Web platform that can’t and shouldn’t be the sole responsibility of libraries, frameworks or Web sites. imho geospatial information is one of them. And a standard for it doesn’t mean to take away choice, it means to provide a baseline from which to progress.

Although analogies are fraught, I think the analogy of two-bladed plugs is a good one. I don’t want to have to mess about with neutral, hot and ground, even though I might know how to, or even want to learn how, in order to charge my phone or turn on the lights. Two-bladed plugs are a standard where I live.