In some cases exons form distinct subregions of the gene, called domains, and at some points in the editing of the primary transcript there may be a choice of exons. Which one to use is then dependent on environmental conditions in the cell. That is, one gene can make several different proteins.

For example a B cell, a kind of lymphocyte, produces antibodies. Early on in its differentiation it needs to retain them at the cell membrane, but later it needs them to circulate, and it achieves this by switching exons, so that the "same" gene now codes for slightly different proteins.

Apparently the latest Human Genome Project figures show that as much as 60% of our genome may have splice variants like this!

In such cases the exons might have evolved as distinct genes and become permanently associated (having a single promoter site and a single terminator).

However not all exons represent bounded domains. In these cases they are probably single genes whose integrity has been invaded by junk DNA introns. Since introns usually begin and end with characteristic two-base sequences, presumably natural selection has spared intrusive junk that just happens to fit the code for domain boundaries.

About the other senses of the word, given by Webster, an exon is one of four officers commanding the Yeomen of the Royal Guard, styled corporal in their commissions, and ranking below an ensign. The word, dating from 1767, derives from the French pronunciation of 'exempt'.
And Exon. is actually the abbreviation of the Latin name of the city of Exeter, viz Exonia, and of its corresponding adjective Exoniensis.