SNIP
>
>Jim, perhaps you can clarify the scope of your wish. However, I still don't believe VFP should have any say in the physical layer, even at the local level.

It is NOT 'VFP having a say in the physical layer' at all, local or remote.

It is simply the OS (not any "physical layer") not returning to VFP until it confirms that the FLUSH has indeed completed. The OS will do whatever it does with the physical layer and when it and the HD have finished their 'talk' then the OS can return to VFP to let it continue processing.

As I said earlier, this is a basic tenet of computing and it should indeed happen on every I/O. Since there is a Microsoft document that states flatly that the FLUSH command does NOT operate this way in the more modern OS' (note, this mens that it DID in older ones) then I want the option to have it go back to doing so.

I don't use FLUSH a lot, but when I do I want to know that it has "ENSURED that the data has been written to disk", as documented in the Help.

By the way, I have started another thread to see if we can learn precisely HOW networked file I/O actually happens.

Does this clarify it? If not, let me ask you this...

Would you find it acceptable if random records in your networked application just disappeared, or they over-wrote other existing records randomly? Of course not. So, then, how do you think that the system(s) involved protect you from those things?

cheers

PS Sorry, but I didn't see this in your message until GerryS replied to you and quoted it.