IJ>Hi Michael Douglass, how are you

IJ>23-Jun-94, Michael Douglass wrote to Ed Blackman
  >           Subject: system and user ini

IJ>EB>>All of my programs use OS2.INI to store the location and name of
  >  >>the program INI files (which I obviously make user
configureable), and I
  >  >>always provide an UNINSTAL.CMD that will delete the program INI file
  >  >>and/or the reference to it in OS2.INI.

IJ> MD> Yes, but wouldn't it be just as easy to store the program INI in the
  > MD> program's directory and use argv[0] to retrieve the directory?

IJ>Then the program could not reside on a readonly-disk such as CDROM,
  >write-protected disk, or write-protected hard disk partition.

Good point, never thought of running it from a CD ROM, but then most
programs run from a CDROM install a small number of files anyways, so
the INI could reside there.

IJ>The best solution would be that all program followed the convention
  >that their INI-file would reside a directory pointed to by a single
  >entry in OS2.INI, so the algorithm for locating a program's private
  >INI file would be:

IJ>  Get 'PROGRAM_INI_LOCATION' entry from OS2.INI
  >  If the entry existed
  >    use that directory for storing the private INI-file
  >  else
  >    ask the user where he wants the INI file to be placed.
  >    write a 'PROGRAM_INI_LOCATION' entry to OS2.INI

That's one of the grandest ideas yet!  It would be a definite good idea
for IBM to incorporate this idea into its programming methods and stress
it.  This would bundle all the INIs together in one place...

Michael Douglass
___
 .Mike's Mail Internet:   MICHAEL.DOUGLASS{at}LCHANCE.SAT.TX.US
             
--- Maximus/2 2.01wb