Google Maps has been getting a lot of customization lately. I'm beginning to wonder if a more organized approach is going to be needed, before this spirals out of control (or grinds to a halt under its own weight). Here are some ideas that I hope will get people thinking and, maybe, working, to build a better, more extensible, more usable system of custom maps and data overlays. Of course, if we go too far down this path, we'll simply be creating a custom GIS tool, but there are worse things, right?<!–break–>
There are now a lot of custom maps being interfaced with the Google Maps system. Here's a quick list of the maps I've seen so far:
Custom Maps with Custom Servers
Integrating Other Map Systems
Overlaying Markers on Regular Map
Other Overlays (Data Displays, not really Maps)
Drop me a note if you see others that aren't on the list, and I'll try to keep this up-to-date!
Making This Better
My Google Maps page is rapidly getting cluttered with all kinds of alternate map sources, some of which are only useful in very specific circumstances (Subway maps, for example). It doesn't make much sense for all kinds of map options to be visible where they're not useful, for one thing, and for another, we're shortly going to have all kinds of people running their own image tile servers. What we need is a central clearinghouse for this data and the scripts that describe getting to it. So, in no particular order, some thoughts about what'd be great to see done, by the community and Google:
- Build a central server to provide map referrals – Basically, just a list of map types and the Javascript routines to implement them. Maybe with some additional information, ala an Extension Room sort of thing. Maybe at sourceforge or mygmaps.com, or even at Google.
- Modify Google viewer to automatically search for appropriate maps – Perhaps the first time the Google Maps viewer initializes (from here out I'm just gonna say “GMaps,” okay?), it asks the server for a list of maps relevant to the current display. If you're looking, for example, at the entire US, it'll ask for any maps that provide coverage for maybe more than half of the country. As you move downwards, it'll re-ask periodically, maybe every other zoom level or whenever you've moved too far away from your last source. Or maybe it can just grab the entire list at the beginning of the session and be done with it (unless we end up with thousands of map services).
- Improve the interface to handle lots of map types – Perhaps only “official” Google maps should be listed in the top of the map window (where all the map types currently appear), and other, local, 3rd party maps could be accessed in a pull-down menu. Maybe people with GMail accounts (and thus Google.com customizations) could even select some maps to always ignore, etc.
- Develop better overlay maps – The CTA and MBTA maps are way cool, but lack some of the detail present in the regular maps. And sometimes it helps to be able to see what's around a station, in the satellite photo. On the other hand, the DC and NYC subway maps give you all that detail, but you lose the subway lines connecting all the stations together. What's needed is a better way to overlay more complicated graphics on the Google map. Perhaps some kind of SVG support could be used, though I suspect that'd be browser dependent. The “polyline” format used for the driving directions has been decoded, but I haven't seen anyone working with it. Using that, with different colors and such, to connect subway stations using non-Google markers, would give a very useful overlay, viewable at all zoom levels.
- Support for different markers – GMaps is currently hard-coded to use the standard Google marker. We've been able to force a way around that, but it's not easy and still requires the same size image. It'd be good to create other markers (like the numbered circles used on the CTA map) that don't have “clickable” area around them and don't cast shadows.
- Live, bounded data retrieval – Currently, GMaps fetches all the points to be displayed at once. I'm not sure how it then handles the points but it does seem that when I plot a few hundred of them, my map gets very slow. It'd be nice if the system could periodically fetch points from a server, giving the server the map's current boundaries and zoom level. The server could then respond with a list of only those points on the current map, and even, taking into account the zoom level, combine some points into single (perhaps larger) marker types.
- Support for local site customization – Finally, it'd be great if one could simply include the GMaps widget, with an immediate application of a custom map or overlay, without having to go through the hoops that the gmaps-standalone system requires. This would be great for very specialized local maps or overlays, such as the earthquake map or a city tourism site.
Some of these would be fairly easy to do. Some would be harder, and would require extensive modifications to the GMaps code. But I really think that it'd leave us with an incredibly powerful system, even if it has to eventually be divorced from Google proper. Sort of a fork. Of course, the Google code isn't public domain, and so we can only take this hacking so far. But if we're able to come up with some great applications on our own, especially open methods of communicating maps and overlays, then maybe it'll get integrated into the official Google system eventually. (or, more likely, reimplemented by Google faster and better).
I can think of some possible uses for this technology that I haven't seen done yet:
- Better subway maps, overlaid on all map types at all zoom levels (showing actual transit lines, not just stations)
- Park maps, hiking trails, bike trails
- Property maps (probably best as an overlay)
- Educational annotation – imagine an overlay on, say, Cape Canaveral giving history for each of the launch pads
- Environmental, climate, and weather map overlays
- Political and other overlays – voting districts, census data, etc.
So am I just being over-idealistic, or is this a good idea? Where do we start?
