The kind of information packing that you see in a good map (like the Times Atlas) is nontrivial. Deciding which features to show, how to label them, and how to arrange the labels is a knapsack problem. Plus, it often involves hard-to-quantify local constraints about what’s most important in a given area. (As a trivial example, a pond is a lot more important in the middle of the Sahara than it would be in Bangladesh. There are many more subtle versions of this.)

While it’s certainly possible to come up with good algorithms, and outfits like Google Maps and Bing Maps have, maps are judged by esthetics and practical usability as well as quantitative criteria[0], so the best results are likely to be hand-tuned.

0. Maps as opposed to the underlying GIS data, which is easier to judge solely by whether it’s complete, precise, and accurate.