One of the interesting differences in design between Google Earth and Microsoft's Virtual Earth is that of when and where data gets rendered. In Google Earth, logical information such as roads, borders and labels are all transported to the client directly and then rendered locally. In Microsoft's product, this information is all rendered on the server and transported in the image tiles that make up the final visualization. One can see the differences in the sample view below which shows approximately the same area of Pittsburgh rendered first in Microsoft's Virtual Earth and then in Google Earth. Have a look at the label for I-279 for comparison.
The Google approach, IMHO, is a win as it allows the client to ensure that the symbols are stable during transitions (navigation and zooming). Microsoft's solution can't ensure this and as one moves around symbols can appear and disappear all over the map.
Of course, from an engineering point of view, Microsoft has simplified the client. As they are deploying this product as a plugin to their browser, rather than a dedicated rich client application, this was probably a fair trade off to get the thing out the door - putting all the opportunities on the server side where they can be better managed.
Comments