lane@SUMEX-AIM.STANFORD.EDU (Christopher Lane) (02/16/91)
Why aren't the user defaults database and NetInfo one in the same rather than separate entities? NetInfo is hierarchy of information that is shared by multiple computers at one level and an individual computer at another. Why not have it go right down to the user level and store user defaults? There are several reasons to do so. First, uniformity and fewer special subsystems and programmer library interfaces. More important though, just as system administrators can set 'fstab' entries for a cluster of machines and have 'fstab' variations on the individual machines, they would be able to set default 'defaults' for all users, that more knowledgeable users could then modify if desired. Thus one could set 'News/NewsHost' globally for *all* users and still allow for individual variations. The NetInfoManger could then be used as a graphic way of reviewing and modifying entries in the individual, machine and network defaults domains. This would be compatible with 'dwrite' and 'dread' which would become alternate access functions to NetInfo. It would not require changing application code as NXGetDefaultValue(), etc. would also become alternate, data-specific access routines to NetInfo. And the user-specific NetInfo data could still live in the individual ~/.NeXT directories. - Christopher -------