The SAVE file is a human readable transcript of the exact commands that you typed and moused to create the cell in the first place. You can edit this file (carefully) and the textual changes will take effect next time you start a new Piglet and read in the definition. The "<cellname>.d" file contains the definition of the cell <cellname>. If <cellname> includes other cells, they will only be called by name in <cellname>.d and will not be defined unless the program can find the corresponding ".d" files. In other words, if you mail a friend a ".d" file, they will be unable to reconstruct your cell if it has any hierarchy. If you want a full, self-contained definition of your work, then you should use the "ARCHIVE" command which defines every cell used in a given drawing.
If there have been any NONAME instances created with the WRAP command, they will be automatically smashed when the device is written to disk. The wrapped instances will still remain in memory, however, so a given device may be written multiple times without destroying the temporary hierarchy created with the wrap command. When the editor is finally closed, all NONAME instances will be permanently lost.
Most commands may be terminated by typing either an EOC or the name of another command. SAVE is an exception to this rule. In the event that the user accidently types the SAVE command more than once, Piglet will quietly ignore the extra SAVE commands. This prevents a command like "SAVE SAVE <a_new_cellname> EXI" from erroneously saving to the current cell in addition to <a_new_cellname>.d.
The PD Piglet homepage is at http://www.omnisterra.com/walker/linux/piglet/intro.htm.