Some background: web developer, with some experience of Google Maps, currently developing a fairly complicated tool for archeological studies. Outside of Google Maps, I have little to no knowledge of mapping services, gis and whatnot.
I have been looking into OpenData sets to implement, and one which I am particularly interested in raised several questions.
The data is at http://www.datiopen.it/it/opendata/Regione_Lazio_Carta_geologica_scala_1_25000 , tab "Mappa".
My first reaction was, geological data, just the thing, awesome.
Then I saw it runs up to about 16,000 elements, all polys - less awesome, but not yet necessarily fatal.
I finally downloaded the data, and found it is a huge shapefile (35Mb zipped - 55Mb unzipped). I learned about gdal just about enough to convert the shapefile to an even worse 112Mb kml.
Even stripping away all the extended data from the kml, or picturing it conveniently stored in a database table, it does not look like any amount of data you would wish to serve over an internet connection.
The linked page, however, does it in but a few seconds.
The way it reacts to mouse events makes me think it simply finds out which poly the clicked coordinates fall within, but, here's what puzzles me, at the same time the contours of the clickable areas are already visible on the map before clicking - suggesting the polys are already loaded.
As this is making no sense to me, I suspect Something Everyone Already Knows But I Don't is at work here, possibly starting with the environment. I have spent some time reading forums and SE itself, but lack of experience leads to lack of perspective.
Can anyone tell by looks what is happening on the linked page, and tell me where to direct my research?