>Am I missing something? Shouldn't directory access be the OS's responsibility? >Particularly for NT, the user should be a member of a domain or user group >that has access to the shared resource, IMHO.

>I guess this is related to my other topic about robust software and >networking, but I think that if you design an app with a security interface, >you could be looking at rewrites in the future due to network changes.

Hi:

I agree it should be the OS's responsibility.
As I said, if I tell WIN95 to access the shared resource in some way, then
VFP5 will be able to access it.

So this was my workaround.
In the same DLL I made, where I establish the connection, I force WIN95 to
access the shared resource, before returning control to VFP5, using
the following WIN32 API function:

DWORD WNetAddConnection2(
LPNETRESOURCE lpNetResource, // points to structure that specifies
// connection details
LPCTSTR lpPassword, // points to password string
LPCTSTR lpUsername, // points to user name string
DWORD fdwConnection // set of bit flags that specify connection
// options
);

Basically it's like telling WIN95 to make a Network Mapping to a
local drive, but in the lpNetResource structure I pass a NULL string
for the local drive. This way there's no visible Network Mapping done but
WIN95 sends the username and password to NT!

It worked perfect.
Hope this can help someone else.