>to the thread on threshholds some comment and/or experience regarding the non-business WinOS 'family'?
>
>I left that message of yours alone in hopes that you can UPDATE it.
>
>That reply itself is great! But I was also (really) after Win9x/WinME information because the "windows networking" in those OSs **is** used with apparent success with VFP apps.
>I could not locate information as regards design intention/limits for those, blaming entirely myself for search incompetence.
>My hope was, and is, to show that, while those OSs may let VFP work for a few users, it would be for no more than N PCs and would would have other limits, probably as regards TRANSACTIONs and as regards amount of buffered updates and the like. Being able to point to something authoritative might deter people from trying to do too much with those OSs and VFP or any other 'database' software.
>
>thanks for what you've said, regardless.

Don't beat yourself up about the search results. There is no way MS is going to admit W9x is less than optimally reliable as an SMB file server. They may not even admit that NT/W2K/XP workstations are more reliable as "Servers" than W9x, even though their underpinnings are the same as the equivalent server versions which they would probably like to (do?) sell as more reliable.

IAC the answer to your question is going to be a spectrum. A "perfect" network of W95 machines may support a lot of users with no problems at all. OTOH 3 poorly maintained Heinz 57 boxes on the same electrical circuit as the arc welder next door will probably blow up 6 times a day.

As SMB servers, I would guess W9x supports all the necessary VFP stuff, like buffering and transactions. It's going to be more a case of "practical" limits, rather than "theoretical" limits.

Maybe you should start another thread - poll users for the maximum # of machines they run peer-to-peer "reliably". You'll probably need to differentiate between occasional/query use vs. heads-down transaction entry.