Hello,<br>I'm recently contemplating the extent with which google maps can draw "markers" on their maps<br><br>if you go to <a href="http://maps.google.com/" target="_blank">maps.google.com</a>,
select "wikipedia" from their layer switcher it draws an immense number
of markers on the screen (I've included a snapshot of this in this
email)<br>
<br>So I wanted to dig deeper in to how they were doing this. I
introspected the requests using firebug and saw that these "markers" are
nothing more than WMS requests, so all I am seeing are pre-rendered
markers, ok fine enough.<br>
<br>but then... I hover over one of the markers... lo and behold the mouse changes! right away, <b>without</b>
another request going back to their server ! so there is more to their
implementation of markers than just simply returning images... <br>
<br>I know they are images because I introspected the requests coming back, they are 256X256 PNG's, one such request is this one<br><br><a href="http://mt1.google.com/mapslt?lyrs=org.wikipedia.en&x=3&y=2&z=4&w=256&h=256&gl=us&hl=en" target="_blank">http://mt1.google.com/mapslt?lyrs=org.wikipedia.en&x=3&y=2&z=4&w=256&h=256&gl=us&hl=en</a><br>
<br>it seems like they return enough information such that as you hover
your mouse over the "marker layer" they know when to change the mouse
over to the pointy finger cursor.. but how do they do that? Do they
return some JSON-like information <br>
that is essentially a map of small rectangles along with some auxiliary
information? then as you hover your mouse they calculate on the fly if
the mouse is <b>within</b> any given rectangle and if it is extract the auxiliary information that came <br>
with that rectangle and display it?<br><br>The data they are returning
has to be more than just a dot, it has to be a rectangle because the "W"
boxes are of varying sizes and simply hovering over to the edge of any
given box the mouse changes and the information that came with that box
comes up.<br>
<br>Anyone ever pondered and/or prototyped such a solution in OpenLayers
as they are providing? If you were to take up such a challenge would
you implement it as I've described it? Is there a much faster/better
solution?<br clear="all"><br>-- <br>Signed,<br>Alessandro Ferrucci<br>