>>The distance between two zips would be arbitrary since it would depend on where in each zipcode you are measuring from. However, if you wanted to give the distance from the center of each zip, that doesn't seem like it would be very hard. You would need a table with zipcodes and their longitude/latitude coordinate quad (a long/lat pair for each corner of the zip region). With these you could easly calculate a long/lat point in the center of the zipcode. Once you have the two zipcode centers, you could calculate the distance between them. This would be a close approximation and has one major drawback: the measurement will only be accurate from the center of each zip. As such, it would have an inherent percent error which is inversely proportional to the distance between the two zips.
>
>Ken Morehouse replied that there is such a table available, and it's at the 5-digit level, thus error would be quite small. The only drawback is the cost: $495 :((, plus $195 per update (since some zips change fairly often, annual updates are advisable)...But that's not unreasonable, considering the work involved putting such a table together...

If you don't want to shell out the cash, I'm sure you can find all the info by doing a little digging on the internet. I have a table with zipcodes/city names I'll be happy to send you if you can use it. Let me know.