I usually recommend that people create mapped drives on XP desktop computers. Mapped drives were popular in Windows 3.x/9x days when networks were small and admins wanted to make it easy for users to find and save stuff on shared folders on a file server. Using a mapped drive, a user could save a file to X: drive instead of having to remember the UNC path (e.g. \\myserver\sales) for the share. Trouble is, if you have a mapped drive but the underlying share is unavailable for some reason (server down, network problem) then users can experience frustrating delays during startup. That’s because if you have any mapped drives defined, XP tries to connect to these drives during startup/logon. And the problem with mapped drives is that once users catch on to their usefulness they tend to proliferate. Then if you alter the underlying share structure on your network, users may end up with drives mapped to nothing. That can also lead to delays in starting some network apps as well.
So my recommendation is to avoid using mapped drives completely. If your users are on an Active Directory network, publish your shares to AD and teach users how to search AD for shares instead using My Network Places.